Matlab Oop

MATLAB Object-Oriented Programming
R2013b
How to Contact MathWorks
Web Newsgroup www.mathworks.com/contact_TS.html Technical Support

www.mathworks.com comp.soft-sys.matlab [email protected] [email protected] [email protected] [email protected] [email protected]
Product enhancement suggestions Bug reports Documentation error reports Order status, license renewals, passcodes Sales, pricing, and general information
508-647-7000 (Phone) 508-647-7001 (Fax) The MathWorks, Inc. 3 Apple Hill Drive Natick, MA 01760-2098
For contact information about worldwide offices, see the MathWorks Web site. Object-Oriented Programming COPYRIGHT 19842013 by The MathWorks, Inc.
The software described in this document is furnished under a license agreement. The software may be used or copied only under the terms of the license agreement. No part of this manual may be photocopied or reproduced in any form without prior written consent from The MathWorks, Inc. FEDERAL ACQUISITION: This provision applies to all acquisitions of the Program and Documentation by, for, or through the federal government of the United States. By accepting delivery of the Program or Documentation, the government hereby agrees that this software or documentation qualifies as commercial computer software or commercial computer software documentation as such terms are used or defined in FAR 12.212, DFARS Part 227.72, and DFARS 252.227-7014. Accordingly, the terms and conditions of this Agreement and only those rights specified in this Agreement, shall pertain to and govern the use, modification, reproduction, release, performance, display, and disclosure of the Program and Documentation by the federal government (or other entity acquiring for or through the federal government) and shall supersede any conflicting contractual terms or conditions. If this License fails to meet the governments needs or is inconsistent in any respect with federal procurement law, the government agrees to return the Program and Documentation, unused, to The MathWorks, Inc.
Trademarks
MATLAB and Simulink are registered trademarks of The MathWorks, Inc. See www.mathworks.com/trademarks for a list of additional trademarks. Other product or brand names may be trademarks or registered trademarks of their respective holders.
Patents
MathWorks products are protected by one or more U.S. patents. Please see www.mathworks.com/patents for more information.
Revision History
March 2008 October 2008 March 2009 September 2009 March 2010 September 2010 April 2011 September 2011 March 2012 September 2012 March 2013 September 2013
Online Online Online Online Online Online Online Online Online Online Online Online
only only only only only only only only only only only only
New for MATLAB 7.6 (Release 2008a) Revised for MATLAB 7.7 (Release 2008b) Revised for MATLAB 7.8 (Release 2009a) Revised for MATLAB 7.9 (Release 2009b) Revised for MATLAB 7.10 (Release 2010a) Revised for Version 7.11 (Release 2010b) Revised for Version 7.12 (Release 2011a) Revised for Version 7.13 (Release 2011b) Revised for Version 7.14 (Release 2012a) Revised for Version 8.0 (Release 2012b) Revised for Version 8.1 (Release 2013a) Revised for Version 8.2 (Release 2013b)
1
Using Object-Oriented Design in MATLAB
Begin Using Object-Oriented Programming on page 1-2 Why Use Object-Oriented Design on page 1-4 Class Diagram Notation on page 1-17 Handle Objects on page 1-19
Begin Using Object-Oriented Programming

In this section... Video Demo of MATLAB Classes on page 1-2 MATLAB Programmer Without Object-Oriented Programming Experience on page 1-2 MATLAB Programmer with Object-Oriented Programming Experience on page 1-2
Video Demo of MATLAB Classes

You can watch a brief presentation on MATLAB class development by clicking this link: Play video
MATLAB Programmer Without Object-Oriented Programming Experience

If you create MATLAB programs, but are not defining classes to accomplish your tasks, start with the following sections: Why Use Object-Oriented Design on page 1-4 Classes in the MATLAB Language on page 2-2 Introductory Examples on page 2-6 Learning Object-Oriented Programming on page 2-7
MATLAB Programmer with Object-Oriented Programming Experience

If have experience with both MATLAB programming and object-oriented techniques, start with the following sections: Class Files on page 3-2 Compatibility with Previous Versions on page 3-46
1-2
Begin Using Object-Oriented Programming
Comparing MATLAB with Other OO Languages on page 3-50
1-3
Why Use Object-Oriented Design

In this section... Approaches to Writing MATLAB Programs on page 1-4 When Should You Start Creating Object-Oriented Programs on page 1-9
Approaches to Writing MATLAB Programs

Creating software applications typically involves designing how to represent the application data and determining how to implement operations performed on that data. Procedural programs pass data to functions, which perform the necessary operations on the data. Object-oriented software encapsulates data and operations in objects that interact with each other via the objects interface. The MATLAB language enables you to create programs using both procedural and object-oriented techniques and to use objects and ordinary functions in your programs.
Procedural Program Design

In procedural programming, your design focuses on steps that must be executed to achieve a desired state. You typically represent data as individual variables or fields of a structure and implement operations as functions that take the variables as arguments. Programs usually call a sequence of functions, each one of which is passed data, and then returns modified data. Each function performs an operation or perhaps many operations on the data.
Object-Oriented Program Design

The object-oriented program design involves: Identifying the components of the system or application that you want to build Analyzing and identifying patterns to determine what components are used repeatedly or share characteristics Classifying components based on similarities and differences
1-4
After performing this analysis, you define classes that describe the objects your application uses.
Classes and Objects

A class describes a set of objects with common characteristics. Objects are specific instances of a class. The values contained in an objects properties are what make an object different from other objects of the same class (an object of class double might have a value of 5). The functions defined by the class (called methods) are what implement object behaviors that are common to all objects of a class (you can add two doubles regardless of their values).
Using Objects in MATLAB Programs

The MATLAB language defines objects that are designed for use in any MATLAB code. For example, consider the try/catch programming construct. If the code executed in the try block generates an error, program control passes to the code in the catch block. This behavior enables your program to provide special error handling that is more appropriate to your particular application. However, you must have enough information about the error to take the appropriate action. MATLAB provides detailed information about the error by passing an MException object to functions executing the try/catch blocks. The following try/catch blocks display the error message stored in an MException object when a function (surf in this case) is called without the necessary arguments:
try surf catch ME disp(ME.message) end Not enough input arguments.
In this code, ME is an object of the MException class, which is returned by the catch statement to the functions workspace. Displaying the value of the objects message property returns information about the error (the surf
1-5
function requires input arguments). However, this is not all the information available in the MException object. You can list the public properties of an object with the properties function:
properties(ME) Properties for class MException: identifier message cause stack
Objects Organize Data

The information returned in an MException object is stored in properties, which are much like structure fields. You reference a property using dot notation, as in ME.message. This reference returns the value of the property. For example,
class(ME.message) ans = char
shows that the value of the message property is an array of class char (a text string). The stack property contains a MATLAB struct:
ME.stack ans = file: [1x90 char] name: 'surf' line: 50
You can simply treat the property reference, ME.stack as a structure and reference its fields:
ME.stack.file ans =
1-6
D:\myMATLAB\matlab\toolbox\matlab\graph3d\surf.m
The file field of the struct contained in the stack property is a character array:
class(ME.stack.file) ans = char
You could, for example, use a property reference in MATLAB functions:

strcmp(ME.stack.name,'surf') ans = 1
Object properties can contain any class of value and can even determine their value dynamically. This provides more flexibility than a structure and is easier to investigate than a cell array, which lacks fieldnames and requires indexing into various cells using array dimensions.
Objects Manage Their Own Data

You could write a function that generates a report from the data returned by MException object properties. This function could become quite complicated because it would have to be able to handle all possible errors. Perhaps you would use different functions for different try/catch blocks in your program. If the data returned by the error object needed to change, you would have to update the functions you have written to use the new data. Objects provide an advantage in that objects define their own operations. A requirement of the MException object is that it can generate its own report. The methods that implement an objects operations are part of the object definition (i.e., specified by the class that defines the object). The object definition might be modified many times, but the interface your program (and other programs) use does not change. Think of your program as a client of the object, which isolates your code from the objects code. To see what methods exist for MException objects, use the methods function:
methods(ME)
1-7
Methods for class MException: addCause eq Static methods: last getReport isequal ne rethrow throw throwAsCaller
You can use these methods like any other MATLAB statement when there is an MException object in the workspace. For example:
ME.getReport ans = Error using ==> surf Not enough input arguments.
Objects often have methods that overload (redefined for the particular class of the object) MATLAB functions (e.g., isequal, fieldnames, etc.). This enables you to use objects just like other values. For example, MException objects have an isequal method. This method enables you to compare these objects in the same way you would compare variables containing doubles. If ME and ME2 are MException objects, you can compare them with this statement:
isequal(ME,ME2)
However, what really happens in this case is MATLAB calls the MException isequal method because you have passed MException objects to isequal. Similarly, the eq method enables you to use the == operator with MException objects:
ME == ME2
Of course, objects should support only those methods that make sense. For example, it would probably not make sense to multiply MException objects so the MException class does not implement methods to do so.
1-8
When Should You Start Creating Object-Oriented Programs

Objects are well integrated into the MATLAB language, regardless of whether you are writing simple functions, working interactively in the command window, or creating large applications. Simple programming tasks are easily implemented as simple functions, but as the magnitude and complexity of your tasks increase, functions become more complex and difficult to manage. As functions become too large, you might break them into smaller functions and pass data from one to the other. However, as the number of functions becomes large, designing and managing the data passed to functions becomes difficult and error prone. At this point, you should consider moving your MATLAB programming tasks to object-oriented designs.
Understanding a Problem in Terms of Its Objects

Thinking in terms of things or objects is simpler and more natural for some problems. You might think of the nouns in your problem statement as the objects you need to define and the verbs as the operations you must perform. For example, consider performing an analysis of economic institutions. It would be difficult to represent the various institutions as procedures even though they are all actors in the overall economy. Consider banks, mortgage companies, credit unions. You can represent each institution as an object that performs certain actions and contains certain data. The process of designing the objects involves identifying the characteristics of these institutions that are important to your application. Identify Commonalities. All of these institutions belong in the general class of lending institutions, so all objects might provide a loan operation and have a Rate property that stores the current interest rate. Identify Differences. You must also consider how each institution differs. A mortgage company might provide only home mortgage loans. Therefore, the loan operation might need be specialized for mortgage companies to provide fixRateLoan and varRateLoan methods to accommodate two loan types.
1-9
Consider Interactions. Institutions can interact, as well. For example, a mortgage company might sell a mortgage to a bank. To support this activity, the mortgage company object would support a sellMortgage operation and the bank object would support a buyMortgage operation. You might also define a loan object, which would represent a particular loan. It might need Amount, Rate, and Lender properties. When the loan is sold to another institution, the Lender property could be changed, but all other information is neatly packaged within the loan object. Add Only What Is Necessary. It is likely that these institutions engage in many activities that are not of interest to your application. During the design phase, you need to determine what operations and data an object needs to contain based on your problem definition. Managing Data. Objects encapsulate the model of what the object represents. If the object represents a kind of lending institution, all the behaviors of lending institutions that are necessary for your application are contained by this object. This approach simplifies the management of data that is necessary in a typical procedural program.
Objects Manage Internal State

In the simplest sense, objects are data structures that encapsulate some internal state, which you access via its methods. When you invoke a method, it is the object that determines exactly what code to execute. In fact, two objects of the same class might execute different code paths for the same method invocation because their internal state is different. The internal workings of the object need not be of concern to your program you simply use the interface the object provides. Hiding the internal state from general access leads to more robust code. If a loan objects Lender property can be changed only by the objects newLender method, then inadvertent access is less likely than if the loan data were stored in a cell array where an indexing assignment statement could damage the data. Objects provide a number of useful features not available from structures and cell arrays. For example, objects provide the ability to:
1-10
Constrain the data assigned to any given property by executing a function to test values whenever an assignment is made Calculate the value of a property only when it is queried and thereby avoid storing data that might be dependent on the state of other data Broadcast notices when any property value is queried or changed, to which any number of listeners can respond by executing functions Restrict access to properties and methods
Reducing Redundancy
As the complexity of your program increases, the benefits of an object-oriented design become more apparent. For example, suppose you need to implement the following procedure as part of your application:
1 Check inputs 2 Perform computation on the first input argument 3 Transform the result of step 2 based on the second input argument 4 Check validity of outputs and return values
This simple procedure is easily implemented as an ordinary function. But now suppose you need to use this procedure again somewhere in your application, except that step 2 must perform a different computation. You could simply copy and paste the first implementation, and then rewrite step 2. Or you could create a function that accepted an option indicating which computation to make, and so on. However, these options lead to more and more complicated code. An object-oriented design could result in a simpler solution by factoring out the common code into what is called a base class. The base class would define the algorithm used and implement whatever is common to all cases that use this code. Step 2 could be defined syntactically, but not implemented, leaving the specialized implementation to the classes that you then derive from this base class.
Step 1 function checkInputs()
1-11
% actual implementation end Step 2 function results = computeOnFirstArg() % specify syntax only end Step 3 function transformResults() % actual implementation end Step 4 function out = checkOutputs() % actual implementation end
The code in the base class is not copied or modified, it is inherited by the various classes you derive from the base class. This reduces the amount of code to be tested, and isolates your program from changes to the basic procedure.
Defining Consistent Interfaces

The use of a class as the basis for similar, but more specialized classes is a useful technique in object-oriented programming. This class is often called an interface class. Incorporating this kind of class into your program design enables you to: Identify the requirements of a particular objective Encode these requirements into your program as an interface class For example, suppose you are creating an object to return information about errors that occur during the execution of specific blocks of code. There might be functions that return special types of information that you want to include in an error report only when the error is generated by these functions. The interface class, from which all error objects are derived, could specify that all error objects must support a getReport method, but not specify how to
1-12
implement that method. The class of error object created for the functions returning special information could implement its version of the getReport method to handle the different data. The requirement defined by the interface class is that all error objects be able to display an error report. All programs that use this feature can rely on it being implement in a consistent way. All of the classes derived from the interface class can create a method called getReport without any name conflicts because it is the class of the object that determines which getReport is called.
Reducing Complexity
Objects reduce complexity by reducing what you need to know to use a component or system. This happens in a couple of ways: Objects provide an interface that hides implementation details. Objects enforce rules that control how objects interact. To illustrate these advantages, consider the implementation of a data structure called a doubly linked list. See Class to Implement Linked Lists on page 2-31 for the actually implementation. Here is a diagram of a three-element list:
n1 Properties Next Prev n2.Prev
n2 Properties Next Prev n2
n3 Properties Next Prev n2.Next
To add a new node to the list, it is necessary to disconnect the existing nodes in the list, insert the new node, and reconnect the nodes appropriately. Here are the basic steps: First disconnect the nodes:
1-13
1 Unlink n2.Prev from n1 2 Unlink n1.Next from n2
Now create the new node, connect it, and renumber the original nodes:
3 Link new.Prev to n1 4 Link new.Next to n3 (was n2) 5 Link n1.Next to new (will be n2) 6 Link n3.Prev to new (will be n2)
n1 Properties Next Prev
Newly inserted node

The details of how methods perform these steps are encapsulated in the class design. Each node object contains the functionality to insert itself into or remove itself from the list. For example, in this class, every node object has an insertAfter method. To add a new node to a list, create the node object and then call its insertAfter method:
nnew = NodeConstructor; nnew.insertAfter(n1)
Because the node class defines the code that implements these operations, this code is: Implemented in an optimal way by the class author Always up to date with the current version of the class
1-14
Well tested Can automatically update old-versions of the objects when they are loaded from MAT-files. The object methods enforce the rules for how the nodes interact. This design removes the responsibility for enforcing rules from the applications that use the objects. It also means the application is less likely to generate errors in its own implementation of the process.
Fostering Modularity
As you decompose a system into objects (car > engine > fuel system > oxygen sensor), you form modules around natural boundaries. These objects provide interfaces by which they interact with other modules (which might be other objects or functions). Often the data and operations behind the interface are hidden from other modules to segregate implementation from interface. Classes provide three levels of control over code modularity: Public Any code can access this particular property or call this method. Protected Only the objects own methods and those of the objects whose class has been derived from this objects class can access this property or call this method. Private Only the objects own methods can access this property or call this method.
Overloaded Functions and Operators

When you define a class, you can overload existing MATLAB functions to work with your new object. For example, the MATLAB serial port class overloads the fread function to read data from the device connected to the port represented by this object. You can define various operations, such as equality (eq) or addition (plus), for a class you have defined to represent your data.
Reduce Code Redundancy

Suppose your application requires a number of dialog windows to interact with users. By defining a class containing all the common aspects of the
1-15
dialog windows, and then deriving the specific dialog classes from this base class, you can: Reuse code that is common to all dialog window implementations Reduce code testing effort due to common code Provide a common interface to dialog developers Enforce a consistent look and feel Apply global changes to all dialog windows more easily
Learning More
See Classes in the MATLAB Language on page 2-2 to learn more about writing object-oriented MATLAB programs.
1-16
Class Diagram Notation
Class Diagram Notation

The diagrams representing classes that appear in this documentation follow the conventions described in the following legend.
1-17
Concept
Graphical representation
Example BankAccount
Object
Properties AccountNumber AccountBalance
Employee Class Properties Name Address
Asset
is_a Stock
FileReader (aggregation) has_a Car (composition)
FileID
Tire
1-18
Handle Objects
Handle Objects
In this section... What Is a Handle Object? on page 1-19 Copying Handles on page 1-19 Modifying Handle Objects in Functions on page 1-20 How to Know if an Object Is a Handle on page 1-22 Deleted Handle Objects on page 1-23
What Is a Handle Object?

Certain kinds of MATLAB objects are handles. When a variable holds a handle, it actually holds a reference to the object. Handle objects enable more than one variable to refer to the same information. Handle-object behavior affects what happens when you copy handle objects and when you pass them to functions.
Copying Handles
All copies of a handle object variable refer to the same underlying object. This means that if h identifies a handle object, then,
h2 = h;
Creates another variable, h2, that refers to the same object as h. For example, the MATLAB audioplayer function creates a handle object that contains the audio source data to reproduce a specific sound segment. The variable returned by the audioplayer function identifies the audio data and enables you to access object functions to play the audio. MATLAB software includes audio data that you can load and use to create an audioplayer object. This sample load audio data, creates the audio player, and plays the audio:
load gong Fs y gongSound = audioplayer(y,Fs);
1-19
play(gongSound)
Suppose you copy the gongSound object handle to another variable (gongSound2):
gongSound2 = gongSound;
The variables gongSound and gongSound2 are copies of the same handle and, therefore, refer to the same audio source. Access the audioplayer information using either variable. For example, set the sample rate for the gong audio source by assigning a new value to the SampleRate property. First get the current sample rate and then set a new sample rate:
sr = gongSound.SampleRate sr = 8192 gongSound.SampleRate = sr*2;
You can use gongSound2 to access the same audio source:

gongSound2.SampleRate ans = 16384
Play the gong sound with the new sample rate:

play(gongSound2)
Modifying Handle Objects in Functions

When you pass an argument to a function, the function copies the variable from the workspace in which you call the function into the parameter variable in the functions workspace.
1-20
Handle Objects
Passing an ordinary (nonhandle) variable to a function does not affect the original variable. For example, myFunc modifies a local variable called var, but when the function ends, the local variable var no longer exists:
function myFunc(var) var = var + 1; end
Define a variable and pass it to myfunc:

x = 12; myFunc(x)
The value of x has not changed after executing myFunc(x):

x x = 12
The myFunc function can return the modified value, which you could assign to the same variable name (x) or another variable.
function out = myFunc(var) out = var + 1; end
Modify a value in myfunc:

x = 12; x = myFunc(x) x ans = 13
When the argument is a handle variable, the function copies only the handle, not the object identified by that handle. Both handles (original and local copy) refer to the same object.
1-21
When the function modifies the data referred to by the object handle, those changes are accessible from the handle variable in the calling workspace without the need to return the modified object. For example, the modifySampleRate function changes the audioplayer sample rate:
function modifySampleRate(audioObj,sr) audioObj.SampleRate = sr; end
Create an audioplayer object and pass it to the modifySampleRate function:

load gong Fs y gongSound = audioplayer(y,Fs); gongSound.SampleRate ans = 8192 modifySampleRate(gongSound,16384) gongSound.SampleRate ans = 16384
The modifySampleRate function does not need to return a modified gongSound object because audioplayer objects are handle objects.
How to Know if an Object Is a Handle

Handle objects are members of the handle class. Therefore, you can always identify an object as a handle using the isa function. isa returns logical true (1) when testing for a handle variable:
load gong Fs y00 gongSound = audioplayer(y,Fs); isa(gongSound,'handle')
1-22
Handle Objects
ans = 1
To determine if a variable is a valid handle object, use isa and isvalid:

if isa(gongSound,'handle') && isvalid(gongSound) ... end
Deleted Handle Objects

When a handle object has been deleted, the handle variables that referenced the object can still exist. These variables become invalid because the object they referred to no longer exists. Calling delete on the object removes the object, but does not clear handle variables. For example, create an audioplayer object:
load gong Fs y gongSound = audioplayer(y,Fs);
The output argument, gongSound, is a handle variable. Calling delete deletes the object along with the audio source information it contains:
delete(gongSound)
However, the handle variable still exists:

gongSound gongSound = handle to deleted audioplayer
The whos command shows gongSound as an audioplayer object:

whos Name Fs gongSound Size 1x1 1x1 Bytes 8 104 Class double audioplayer Attributes
1-23
42028x1
336224
double
The handle gongSound no longer refers to a valid object:

isvalid(gongSound) ans = 0
You cannot call functions on the invalid handle variable:

play(gongSound) Invalid or deleted object.
You cannot access properties with the invalid handle variable:

gongSound.SampleRate Invalid or deleted object.
To remove the variable, gongSound, use clear:

clear gongSound whos Name Size Fs y 1x1 42028x1
Bytes 8 336224
Class double double
Attributes
1-24
2
MATLAB Classes Overview
Classes in the MATLAB Language on page 2-2 Detailed Information and Examples on page 2-8 Developing Classes Typical Workflow on page 2-11 Working with Objects in Functions on page 2-18 Class to Represent Structured Data on page 2-22 Class to Implement Linked Lists on page 2-31 Class for Graphing Functions on page 2-44
Classes in the MATLAB Language

In this section... Classes on page 2-2 Some Basic Relationships on page 2-4 Introductory Examples on page 2-6 Learning Object-Oriented Programming on page 2-7
Classes
In the MATLAB language, every value is assigned to a class. For example, creating a variable with an assignment statement constructs a variable of the appropriate class:
>> a = 7; >> b = 'some string'; >> whos Name Size a b 1x1 1x11
Bytes 8 22
Class double char
Basic commands like whos display the class of each value in the workspace. This information helps MATLAB users recognize that some values are characters and display as text while other values might be double, single, or other types of numbers. Some variables can contain different classes of values like cells.
User-Defined Classes
You can create your own MATLAB classes. For example, you could define a class to represent polynomials. This class could define the operations typically associated with MATLAB classes, like addition, subtraction, indexing, displaying in the command window, and so on. However, these operations would need to perform the equivalent of polynomial addition, polynomial subtraction, and so on. For example, when you add two polynomial objects:
p1 + p2
2-2
the plus operation would know how to add polynomial objects because the polynomial class defines this operation. When you define a class, you overload special MATLAB functions (plus.m for the addition operator) that are called by the MATLAB runtime when those operations are applied to an object of your class. See A Polynomial Class on page 17-2 for an example that creates just such a class.
MATLAB Classes Key Terms

MATLAB classes use the following words to describe different parts of a class definition and related concepts. Class definition Description of what is common to every instance of a class. Properties Data storage for class instances Methods Special functions that implement operations that are usually performed only on instances of the class Events Messages that are defined by classes and broadcast by class instances when some specific action occurs Attributes Values that modify the behavior of properties, methods, events, and classes Listeners Objects that respond to a specific event by executing a callback function when the event notice is broadcast Objects Instances of classes, which contain actual data values stored in the objects properties Subclasses Classes that are derived from other classes and that inherit the methods, properties, and events from those classes (subclasses facilitate the reuse of code defined in the superclass from which they are derived). Superclasses Classes that are used as a basis for the creation of more specifically defined classes (i.e., subclasses). Packages Folders that define a scope for class and function naming
2-3
These are general descriptions of these components and concepts. This documentation describes all of these components in detail.
Some Basic Relationships

This section discusses some of the basic concepts used by MATLAB classes.
Classes
A class is a definition that specifies certain characteristics that all instances of the class share. These characteristics are determined by the properties, methods, and events that define the class and the values of attributes that modify the behavior of each of these class components. Class definitions describe how objects of the class are created and destroyed, what data the objects contain, and how you can manipulate this data.
Class Hierarchies
It sometimes makes sense to define a new class in terms of existing classes. This enables you to reuse the designs and techniques in a new class that represents a similar entity. You accomplish this reuse by creating a subclass. A subclass defines objects that are a subset of those defined by the superclass. A subclass is more specific than its superclass and might add new properties, methods, and events to those inherited from the superclass. Mathematical sets can help illustrate the relationships among classes. In the following diagram, the set of Positive Integers is a subset of the set of Integers and a subset of Positive numbers. All three sets are subsets of Real numbers, which is a subset of All Numbers. The definition of Positive Integers requires the additional specification that members of the set be greater than zero. Positive Integers combine the definitions from both Integers and Positives. The resulting subset is more specific, and therefore more narrowly defined, than the supersets, but still shares all the characteristics that define the supersets.
2-4
All Numbers Integers Positive Integers Positives Reals
The is a relationship is a good way to determine if it is appropriate to define a particular subset in terms of existing supersets. For example, each of the following statements makes senses: A Positive Integer is an Integer A Positive Integer is a Positive number If the is a relationship holds, then it is likely you can define a new a class from a class or classes that represent some more general case.
Reusing Solutions
Classes are usually organized into taxonomies to foster code reuse. For example, if you define a class to implement an interface to the serial port of a computer, it would probably be very similar to a class designed to implement an interface to the parallel port. To reuse code, you could define a superclass that contains everything that is common to the two types of ports, and then
2-5
derive subclasses from the superclass in which you implement only what is unique to each specific port. Then the subclasses would inherit all of the common functionality from the superclass.
Objects
A class is like a template for the creation of a specific instance of the class. This instance or object contains actual data for a particular entity that is represented by the class. For example, an instance of a bank account class is an object that represents a specific bank account, with an actual account number and an actual balance. This object has built into it the ability to perform operations defined by the class, such as making deposits to and withdrawals from the account balance. Objects are not just passive data containers. Objects actively manage the data contained by allowing only certain operations to be performed, by hiding data that does not need to be public, and by preventing external clients from misusing data by performing operations for which the object was not designed. Objects even control what happens when they are destroyed.
Encapsulating Information
An important aspect of objects is that you can write software that accesses the information stored in the object via its properties and methods without knowing anything about how that information is stored, or even whether it is stored or calculated when queried. The object isolates code that accesses the object from the internal implementation of methods and properties. You can define classes that hide both data and operations from any methods that are not part of the class. You can then implement whatever interface is most appropriate for the intended use.
Introductory Examples
The following examples illustrate some basic features of MATLAB classes. Developing Classes Typical Workflow on page 2-11 applies object-oriented thinking to a familiar concept to illustrate the process of designing classes.
2-6
Working with Objects in Functions on page 2-18 shows advantages of using objects to define certain operations and how smoothly object fit in a function-oriented workflow. Class to Represent Structured Data on page 2-22 shows the application of object-oriented techniques to managing data. Class to Implement Linked Lists on page 2-31 using a handle class to implement a doubly linked list.
Learning Object-Oriented Programming

The following references can help you develop a basic understanding of object-oriented design and concepts. Shalloway, A., J. R. Trott, Design Patterns Explained A New Perspective on Object-Oriented Design.. Boston, MA: Addison-Wesley 2002. Gamma, E., R. Helm, R. Johnson, J. Vlissides, Design Patterns Elements of Reusable Object-Oriented Software. Boston, MA: Addison-Wesley 1995. Freeman, E., Elisabeth Freeman, Kathy Sierra, Bert Bates, Head First Design Patterns. Sebastopol, CA 2004. See Wikipedia :Object Oriented Programming
2-7
Detailed Information and Examples

Rapid Access to Information
This section provides a gateway to both conceptual information and example implementations. It enables you to scan the information available for broad topics Background Information and Discussion Attribute Tables List of all class member attributes: Attribute Tables Classes in the MATLAB Language on page 2-2 for an introduction to object-oriented programming concepts. User-Defined Classes on page 4-2 for an overview of classes features. Developing Classes Typical Workflow on page 2-11 for a simple example Class to Represent Structured Data on page 2-22 Class to Implement Linked Lists on page 2-31 A Polynomial Class on page 17-2 A Simple Class Hierarchy on page 18-2 Containing Assets in a Portfolio on page 18-19 Attributes Class Attributes on page 4-6 for a list of class attributes Hierarchies of Classes Concepts on page 10-2 describes how classes can be built on other classes A Simple Class Hierarchy on page 18-2
Topic Attributes (all) Classes
Code Examples
2-8
Detailed Information and Examples
(Continued) Background Information and Discussion Attribute Tables Creating Subclasses Syntax and Techniques on page 10-7 Modifying Superclass Methods and Properties on page 10-14 Kinds of classes Comparing Handle and Value Classes on page 5-2 The Handle Superclass on page 5-11 a detailed description of the abstract class. Properties Defining Properties on page 6-5 for an overview of what properties are and how to use them Property Definition Block on page 6-5 shows how to specify initial values Attributes Specifying Property Attributes on page 6-7 for a list of property attributes Dynamic Properties Adding Properties to an Instance on page 6-27 Methods How to Use Methods on page 7-2 for an overview of methods Attributes Method Attributes on page 7-5 for a list of method attributes Restricting Properties to Specific Values on page 2-25 Using a Dependent Property on page 2-27 Assigning Data to the Dynamic Property on page 6-29 Class to Implement Linked Lists on page 2-31 Specializing the dlnode Class on page 2-40
Topic Attributes (all)
Code Examples
2-9
(Continued) Background Information and Discussion Attribute Tables Class Constructor Methods on page 7-16 for information about constructor methods Handle Class Destructor on page 5-16 Property Access Methods on page 6-14 Implementing a Set/Get Interface for Properties on page 5-23 Events Events and Listeners Concepts on page 9-11 for an overview of how events work Events and Listeners Syntax and Techniques on page 9-18 for the syntax used to define events and listeners Update Graphs Using Events and Listeners on page 9-35 for a complete example that uses events and listeners, including a property listener Restricting Properties to Specific Values on page 2-25 Simplifying the Interface with a Constructor on page 2-26
Topic Attributes (all)
Code Examples
2-10
Developing Classes Typical Workflow

In this section... Formulating a Class on page 2-11 Implementing the BankAccount Class on page 2-13 Implementing the AccountManager Class on page 2-15 Using the BankAccount Class on page 2-16
Formulating a Class
This example discusses the design and implementation of a simple class. To design a class that represents a bank account, first determine the elements of data and the operations that form your abstraction of a bank account. For example, a bank account has: An account number An account balance A current status (open, closed, etc.) You need to perform certain operations on a bank account: Deposit money Withdraw money You might also want the bank account to send a notice if the balance is too low and an attempt is made to withdraw money. When this event occurs, the bank account can broadcast a notice to other entities that are designed to listen for these notices, such as an account manager program. The account manager program can take action in response to the event. In this class, the status of all bank accounts is determined by an account manager program that looks at the account balance and assigns one of three values: open Account balance is a positive value
2-11
overdrawn Account balance is overdrawn, but by $200 or less. closed Account balance is overdrawn by more than $200. MATLAB classes store data in properties, implement operations with methods, and support notifications with events and listeners. Therefore, the bank account class needs to implement these components, which are discussed in the following sections.
Class Data
The class needs to define these properties to store the account number, account balance, and the account status: AccountNumber MATLAB assigns a value to this property when you create an instance of the class. AccountBalance The class operation of depositing and withdrawing money assigns values to this property. AccountStatus MATLAB sets this property to an initial value when an instance of the class is created. It is then changed by methods from the AccountManager class whenever the value of the AccountBalance falls below 0. The first two properties contain information that only the class can change, so the SetAccess attribute is set to private (only class methods can set these values). An external program sets the value of the AccountStatus property. This program needs access to the property, so the propertys SetAccess attribute is left as public (any code can access this property value).
Class Operations
There are three operations that the class must be able to perform, so there needs to be three methods: deposit Update the AccountBalance property when a deposit transaction occurs
2-12
withdraw Update the AccountBalance property when a withdrawal transaction occurs BankAccount Create an initialized instance of the class
Class Events
The account manager program changes the status of bank accounts having negative balances. To implement this action, the BankAccount class triggers an event when a withdrawal results in a negative balance. Therefore, the triggering of the InsufficientsFunds event occurs from within the withdraw method. To define an event, specify a name within an events block. Trigger the event by a call to the notify handle class method. Because InsufficientsFunds is not a predefined event, you can name it with any string and trigger it with any action.
Implementing the BankAccount Class

It makes sense for there to be only one set of data associated with any instance of a BankAccount class. You would not want independent copies of the object that could have, for example, different values for the account balance. Therefore, the BankAccount class should be implemented as a handle class. All copies of a given handle object refer to the same data.
Commented Example Code

You can display the code for this example in a popup window that contains detailed comments and links to related sections of the documentation: BankAccount class AccountManager class Open both class files in your editor by clicking this link: Open in editor
2-13
Class Definition
classdef BankAccount < handle properties (Hidden) AccountStatus = 'open'; end % The following properties can be set only by class methods properties (SetAccess = private) AccountNumber AccountBalance = 0; end % Define an event called InsufficientFunds events InsufficientFunds end methods function BA = BankAccount(AccountNumber,InitialBalance) BA.AccountNumber = AccountNumber; BA.AccountBalance = InitialBalance; % Calling a static method requires the class name % addAccount registers the InsufficientFunds listener on this instance AccountManager.addAccount(BA); end function deposit(BA,amt) BA.AccountBalance = BA.AccountBalance + amt; if BA.AccountBalance > 0 BA.AccountStatus = 'open'; end end function withdraw(BA,amt) if (strcmp(BA.AccountStatus,'closed')&& BA.AccountBalance < 0) disp(['Account ',num2str(BA.AccountNumber),' has been closed.']) return end newbal = BA.AccountBalance - amt; BA.AccountBalance = newbal; % If a withdrawal results in a negative balance, % trigger the InsufficientFunds event using notify if newbal < 0 notify(BA,'InsufficientFunds')
2-14
end end % withdraw end % methods end % classdef
Implementing the AccountManager Class

The AccountManager class provides two methods that implement and register a listener for the InsufficientsFunds event, which is defined for all BankAccount objects. The BankAccount class constructor method calls addAccount to register the listener for the instance being created.
Class Definition
classdef AccountManager methods (Static) function assignStatus(BA) if BA.AccountBalance < 0 if BA.AccountBalance < -200 BA.AccountStatus = 'closed'; else BA.AccountStatus = 'overdrawn'; end end end function addAccount(BA) % Call the handle addlistener method % Object BA is a handle class addlistener(BA, 'InsufficientFunds', ... @(src, evnt)AccountManager.assignStatus(src)); end end end
Note that the AccountManager class is never instantiated. It serves as a container for the event listener used by all BankAccount objects.
2-15
Using the BankAccount Class

The BankAccount class, while overly simple, demonstrates how MATLAB classes behave. For example, create a BankAccount object with a serial number and an initial deposit of $500:
BA = BankAccount(1234567,500); BA.AccountNumber ans = 1234567 BA.AccountBalance ans = 500 BA.AccountStatus ans = open
Now suppose you make a withdrawal of $600, which results in a negative account balance:
BA.withdraw(600) BA.AccountBalance ans = -100 BA.AccountStatus ans = overdrawn
When the $600 withdrawal occurred, the InsufficientsFunds event was triggered. Because the AccountBalance is not less than $200, the AccountStatus was set to overdrawn:
BA.withdraw(200) BA.AccountBalance ans = -300 BA.AccountStatus ans = closed
Now the AccountStatus has been set to closed by the listener and further attempts to make withdrawals are blocked:
2-16
BA.withdraw(100) Account 1234567 has been closed
If the AccountBalance is returned to a positive value by a deposit, then the AccountStatus is returned to open and withdrawals are allowed again:
BA.deposit(700) BA.AccountStatus ans = open BA.withdraw(100) BA.AccountBalance ans = 300
2-17
Working with Objects in Functions

In this section... Flexible Workflow on page 2-18 Performing a Task with an Object on page 2-18 Using Object Functionality on page 2-20
Flexible Workflow
The MATLAB language does not require you to define classes for all the code you write. You can use objects along with ordinary functions. This section illustrates the use of an object that implements the basic task of writing text to a file. Then this object is used in a function to write a text file template for a class definition.
Performing a Task with an Object

One of the advantages of defining a class instead of simply writing a function to perform a task is that classes provide better control over related data. For example, consider the task of writing data to a file. It involves the following steps: Opening a file for writing and saving the file identifier Using the file identifier to write data to the file Using the file identifier to close the file
The Filewriter Class

This simple class definition illustrates how you might create a class to write text to a file. It shows how you can use a class definition to advantage by: Hiding private data The caller does not need to manage the file identifier. Ensuring only one file identifier is in use at any time Copies of handle objects reference the same file identifier as the original. Providing automatic file closing when the object is deleted the objects delete method takes care of cleanup without needing to be called explicitly.
2-18
This class is derived from the handle class so that a Filewriter object is a handle object. All copies of handle objects reference the same internal data so there will be only one file identifier in use, even if you make copies of the object. Also, handle classes define a delete method which is called automatically when a handle object is destroyed. This example overrides the delete method to close the file before the file identifier is lost and the file is left open.
classdef Filewriter < handle % Property data is private to the class properties (SetAccess = private, GetAccess = private) FileID end % properties methods % Construct an object and % save the file ID function obj = Filewriter(filename) obj.FileID = fopen(filename,'a'); end function writeToFile(obj,text_str) fprintf(obj.FileID,'%s\n',text_str); end % Delete methods are always called before a object % of the class is destroyed function delete(obj) fclose(obj.FileID); end end % methods end % class
Using a Filewriter Object

Note that the user provides a file name to create a Filewriter object, and then uses the class writeToFile method to write text to the file. The following statements create a file named mynewclass.m and write one line to it. The clear all command deletes the Filewriter object, which causes its delete method to be called and the file is closed.
>> fw = Filewriter('mynewclass.m');
2-19
>> fw.writeToFile('classdef mynewclass < handle') >> clear fw >> type mynewclass classdef mynewclass < handle
Using Object Functionality

Filewriter objects provide functionality that you can use from functions and within other classes. You can create an ordinary function that uses this object, as the writeClassFile function does below.
This example creates only one simple class template, but another version might accept a cell array of attribute name/value pairs, method names, and so on.
function writeClassFile(classname,superclass) % Use a Filewriter object to write text to a file fw = Filewriter([classname '.m']); if nargin > 1 fw.writeToFile(['classdef ' classname ' < ' superclass]) else fw.writeToFile(['classdef ' classname]) end fw.writeToFile(' properties ') fw.writeToFile(' ') fw.writeToFile(' end % properties') fw.writeToFile(' ') fw.writeToFile(' methods ') fw.writeToFile([' function obj = ' classname '()']) fw.writeToFile(' ') fw.writeToFile(' end') fw.writeToFile(' end % methods') fw.writeToFile('end % classdef') delete(fw) % Delete object, which closes file end
To create a class file template, call writeClassFile with the name of the new class and its superclass. Use the type command to display the contents of the file:
2-20
>> writeClassFile('myNewClass','handle') >> type myNewClass classdef myNewClass < handle properties end % properties methods function obj = myNewClass() end end % methods end % classdef
More Information on These Techniques

The Handle Superclass on page 5-11 Handle Class Destructor on page 5-16
2-21
Class to Represent Structured Data

In this section... Commented Example Code on page 2-22 Objects As Data Structures on page 2-22 Structure of the Data on page 2-23 The TensileData Class on page 2-23 Creating an Instance and Assigning Data on page 2-24 Restricting Properties to Specific Values on page 2-25 Simplifying the Interface with a Constructor on page 2-26 Using a Dependent Property on page 2-27 Displaying TensileData Objects on page 2-28 Method to Plot Stress vs. Strain on page 2-29

Open class code in a popup window Use this link if you want to see the final code for this class annotated with links to descriptive sections. Open class definition file in the MATLAB editor. Use this link if you want to save and modify your version of the class. To use the class, create a folder named @TensileData and save TensileData.m to this folder. The parent folder of @TensileData must be on the MATLAB path.
Objects As Data Structures

This example defines a class for storing data with a specific structure. Using a consistent structure for data storage makes it easier to create functions that operate on the data. While a MATLAB struct with field names describing the particular data element is a useful way to organize data, the use of a class to define both the data storage (properties) and operations you can perform on that data (methods) provides advantages, as this example illustrates.
2-22
Concepts on Which This Example Is Based.

For purposes of this example, the data represents tensile stress/strain measurements, which are used to calculate the elastic modulus of various materials. In simple terms, stress is the force applied to a material and strain is the resulting deformation. Their ratio defines a characteristic of the material. While this is an over simplification of the process, it suffices for this example.
Structure of the Data

The following table describes the structure of the data.
Data
Material SampleNumber Stress Strain Modulus
Description Character string identifying the type of material tested Number of a particular test sample Vector of doubles representing the stress applied to the sample during the test. Vector of doubles representing the strain at the corresponding values of the applied stress. Double defining an elastic modulus of the material under test, which is calculated from the stress and strain data
The TensileData Class

This class is designed to store data, so it defines a property for each of the data elements. The following class block defines five properties and specifies their initial values according to the type of data each will contain. Defining initial values is not required, but can be useful if a property value is not assigned during object creation. Note that this example begins with a simple implementation of the class and builds on this implementation to illustrate how features enhance the usefulness of the class.
2-23
classdef TensileData properties Material = ''; SampleNumber = 0; Stress Strain Modulus = 0; end end
Creating an Instance and Assigning Data

Create a TensileData object and assign data to it with the following statements:
td = TensileData; td.Material = 'Carbon Steel'; td.SampleNumber = 001; td.Stress = [2e4 4e4 6e4 8e4]; td.Strain = [.12 .20 .31 .40]; td.Modulus = mean(td.Stress./td.Strain);
Advantages of a Class vs. a Structure Array

Treat the TensileData object (td in the statements above) much as you would any MATLAB structure array. However, defining a specialized data structure as a class has advantages over using a general-purpose data structure, like a MATLAB struct: Users cannot accidentally misspell a field name without getting an error. For example, typing the following:
>>td.Modulis = ...
would simply add a new field to a structure array, but returns an error when td is an instance of the TensileData class. A class is easy to reuse. Once you have defined the class, you can easily extend it with subclasses that add new properties.
2-24
A class is easy to identify. A class has a name so that you can identify objects with the whos and class functions and the Workspace browser. The class name makes it easy to refer to records with a meaningful name. A class can validate individual field values when assigned, including class or value. A class can restrict access to fields, for example, allowing a particular field to be read, but not changed. The next section describes how to add type checking and how to restrict property access in the TensileData class.
Restricting Properties to Specific Values

Restrict the values to which a property can be set by defining a property set access method. MATLAB software then calls this function whenever a value is set for a property, including when creating the object.
Defining the Material Property Set Function

The property set method restricts the assignment of the Material property to one of the following strings: aluminum, stainless steel, or carbon steel. Add this function definition to the methods block.
classdef TensileData properties Material = 'carbon steel'; SampleNumber = 0; Stress Strain Modulus end % properties methods function obj = set.Material(obj,material) if ~(strcmpi(material,'aluminum') ||... strcmpi(material,'stainless steel') ||... strcmpi(material,'carbon steel')) error('Material must be aluminum, stainless steel, or carbon steel') end
2-25
obj.Material = material; end % set.Material end% methods end% classdef
When an attempt is made to set the Material property, the MATLAB runtime passes the object and the specified value to the propertys set.Material function (the obj and the material input arguments). In this case, if the value does not match the acceptable values, the function returns an error. Otherwise, the specified value is used to set the property. Only the set method can directly access the property in the object (without calling the property set method). For example:
td = TensileData; td.Material = 'composite'; Error using TensileData/set.Material Material must be aluminum, stainless steel, or carbon steel
Simplifying the Interface with a Constructor

You can simplify the interface to the TensileData class by adding a constructor function that: Enables you to pass the data as arguments to the constructor Assigns values to properties The constructor is a method having the same name as the class.
function td = TensileData(material,samplenum,stress,strain) if nargin > 0 % Support calling with 0 arguments td.Material = material; td.SampleNumber = samplenum; td.Stress = stress; td.Strain = strain; end end % TensileData
2-26
Using the constructor, you can create a TensileData object fully populated with data using the following statement:
td = TensileData('carbon steel',1,[2e4 4e4 6e4 8e4],[.12 .20 .31 .40]);
Calculating Modulus
Note that the constructor function does not have an input argument for the value of the Modulus property. This is because the value of the Modulus: Is easy to calculate from the Stress and Strain property values Needs to change if the value of the Stress or Strain property changes Therefore, it is better to calculate the value of the Modulus property only when its value is requested. You can do this with a property get access method, which is described in the next section.
Using a Dependent Property

TensileData objects do not store the value of the Modulus property; instead this value is calculated whenever it is requested. This approach enables you to update the Stress and Strain property data at any time without having to recalculate the value of the Modulus property.
Modulus Property Get Method

The Modulus property depends on Stress and Strain, so its Dependent attribute is set to logical true. To do this, create another properties block to set the Dependent attribute. Also, because the get.Modulus method calculates and returns the value of the Modulus property, you should set the propertys SetAccess attribute to private.
properties (Dependent = true, SetAccess = private) Modulus end
Define the propertys get method in a methods block.
2-27
methods function modulus = get.Modulus(obj) ind = find(obj.Strain > 0); % Find nonzero strain modulus = mean(obj.Stress(ind)./obj.Strain(ind)); end % Modulus get method end % methods
This function simply calculates the average ratio of stress to strain data after eliminating zeros in the denominator data. The MATLAB runtime calls the get.Modulus method when the property is queried. For example,
td = TensileData('carbon steel',1,[2e4 4e4 6e4 8e4],[.12 .20 .31 .40]); td.Modulus ans = 1.9005e+005
Modulus Property Set Method

To set the value of a Dependent property, the class must implement a property set method. There is no need to enable explicit setting of the Modulus property, but a set method enables you to provide a customized error message. The Modulus set method references the current property value and then returns an error:
methods function obj = set.Modulus(obj,~) fprintf('%s%d\n','Modulus is: ',obj.Modulus) error('You cannot set Modulus explicitly'); end % Modulus set method end % methods
Displaying TensileData Objects

The TensileData class can implement a disp method that controls what is displayed when an object of this class is shown on the command line (for example, by an assignment statement not terminated by a semicolon).
2-28
The TensileData disp method displays the value of the Material, SampleNumber, and Modulus properties. It does not display the Stress and Strain property data since these properties contain raw data that is not easily viewed in the command window. The plot method (described in the next section) provides a better way to display stress and strain data. The disp method uses fprintf to display formatted text in the command window:
methods function disp(td) fprintf(1,'Material: %s\nSample Number: %g\nModulus: %1.5g\n',... td.Material,td.SampleNumber,td.Modulus); end % disp end % methods
Method to Plot Stress vs. Strain

It is useful to view a graph of the stress/strain data to determine the behavior of the material over a range of applied tension. A TensileData object contains the stress and strain data so it is useful to define a class method that is designed to plot this data. The TensileData plot method creates a linear graph of the stress versus strain data and adds a title and axis labels to produce a standardized graph for the tensile data records:
function plot(td,varargin) plot(td.Strain,td.Stress,varargin{:}) title(['Stress/Strain plot for Sample',... num2str(td.SampleNumber)]) ylabel('Stress (psi)') xlabel('Strain %') end % plot
The first argument to this method is a TensileData object, which contains the data and is used by the MATLAB runtime to dispatch to the TensileData class plot method and not the built-in plot function. The variable list of arguments that follow are passed directly to the built-in plot function from within the method. This enables the TensileData plot
2-29
method to behave like the built-in plot function, which allows you to pass line specifier arguments or property name/value pairs along with the data. For example, plotting the following object:
td = TensileData('carbon steel',1,[2e4 4e4 6e4 8e4],[.12 .20 .31 .40]); plot(td,'-+g','LineWidth',2)
produces this graph.
2-30
Class to Implement Linked Lists

In this section... Commented Example Code on page 2-31 Important Concepts Demonstrated on page 2-31 dlnode Class Design on page 2-32 Creating Doubly Linked Lists on page 2-33 Why a Handle Class for Linked Lists? on page 2-34 Defining the dlnode Class on page 2-35 Specializing the dlnode Class on page 2-40

Open class code in a popup window Use this link if you want to see the code for this class annotated with links to descriptive sections. Open class definition file in the MATLAB editor. Use this link if you want to save and modify your version of the class. To use the class, create a folder named @dlnode and save dlnode.m to this folder. The parent folder of @dlnode must be on the MATLAB path. Alternatively, save dlnode.m to a path folder.
Important Concepts Demonstrated

This section discusses concepts that are important in object-oriented design, and which are illustrated in this example.
Encapsulation
This example shows how classes encapsulate the internal structure used to implement the class design (a doubly linked lists). Encapsulation conceals the internal workings of the class from other code and provides a stable interface to programs that use this class. It also prevents client code from misusing the class because only class methods can access certain class data.
2-31
Class methods define the operations that you can perform on nodes of this class. These methods hide the potentially confusing process of inserting and removing nodes, while at the same time providing an interface that performs operations simply: Creating a node by passing the constructor a data value Inserting nodes with respect to other nodes in the list (before or after) Removing nodes from the list See Defining the dlnode Class on page 2-35 for the implementation details.
Handle Class Behavior

This example shows an application of a handle class and explains why this is the best choice for the class. See Why a Handle Class for Linked Lists? on page 2-34.
dlnode Class Design

This example defines a class for creating the nodes of doubly linked lists in which each node contains: Data array Link to the next node Link to the previous node Each node has methods that enables the node to be: Disconnected from a linked list Connected before a specified node in a linked list Connected after a specific node in a linked list
Class Properties
The dlnode class implements each node as a handle object with three properties:
2-32
Data Contains the data for this node Next Contains the handle of the next node in the list (SetAccess = private) Prev Contains the handle of the previous node in the list (SetAccess = private) This diagram shows a three-node list n1, n2, and n3. It also shows how the nodes reference the next and previous nodes.
n1 Properties Next Prev n2.Prev
n2 Properties Next Prev n2
n3 Properties Next Prev n2.Next
Class Methods
The dlnode class implements the following methods: dlnode Constructs a node and assigns the value passed as input to the Data property insertAfter Inserts this node after the specified node insertBefore Inserts this node before the specified node disconnect Removes this node from the list disp Overloads default disp function so that the Data property displays on the command line for scalar objects and the dimension of the array displays for object arrays delete Removes this node from the list before it is destroyed
Creating Doubly Linked Lists

Create a node by passing the nodes data to the dlnode class constructor. For example, these statements create three nodes with sequential integer data just for simplicity:
2-33
n1=dlnode(1); n2=dlnode(2); n3=dlnode(3);
Build these nodes into a doubly linked list using the class methods:
n2.insertAfter(n1) n3.insertAfter(n2)
Now the three nodes are linked. The dlnode disp method returns the data for the node referred to:
n1.Next % Points to n2 ans = Doubly-linked list node with data: 2 n2.Next.Prev % Points back to n2 ans = Doubly-linked list node with data: 2 n1.Next.Next % Points to n3 ans = Doubly-linked list node with data: 3 n3.Prev.Prev % Points to n1 ans = Doubly-linked list node with data: 1
Why a Handle Class for Linked Lists?

Each node is unique in that no two nodes can be previous to or next to the same node. Suppose a node object, node, contains in its Next property the handle of the next node object, node.Next. Similarly, the Prev property contains the handle of the previous node, node.Prev. Using the three-node linked list defined in the previous section, you can demonstrate that the following statements are true:
n1.Next == n2 n2.Prev == n1
2-34
Now suppose you assign n2 to x:

x = n2;
The following two equalities are then true:

x == n1.Next x.Prev == n1
But each instance of a node is unique so there is only one node in the list that can satisfy the conditions of being equal to n1.Next and having a Prev property that contains a handle to n1. Therefore, x must point to the same node as n2. This means there has to be a way for multiple variables to refer to the same object. The MATLAB handle class provides a means for both x and n2 to refer to the same node. All instances of the handle class are handles that exhibit the copy behavior described previously. Notice that the handle class defines the eq method (use methods('handle') to list the handle class methods), which enables the use of the == operator with all handle objects. See Comparing Handle and Value Classes on page 5-2 for more information on kinds of MATLAB classes. See The Handle Superclass on page 5-11 for more information about the handle class.
Defining the dlnode Class

The following examples use this doubly linked list (see Commented Example Code on page 2-31 before using this class):
n1 = dlnode(1); n2 = dlnode(2); n3 = dlnode(3); n2.insertAfter(n1) n3.insertAfter(n2)
2-35
Class Properties
The dlnode class is itself a handle class because it is derived from the handle class. Note that only class methods can set the Next and Prev properties (SetAccess = private). Using private set access prevents client code from performing any incorrect operation with these properties. The dlnode class defines methods that perform all the operations that are allowed on these nodes. Here are the property definition blocks:
classdef dlnode < handle properties Data end properties (SetAccess = private) Next Prev end
Creating a Node Object

To create a node object, you need to specify only the nodes data.
function node = dlnode(Data) if nargin > 0 node.Data = Data; end end
When you add the node to a list, the class methods that perform the insertion set the Next and Prev properties. See Inserting Nodes on page 2-38.
Disconnecting Nodes
The disconnect method removes a node from a list and repairs the list by reconnecting the appropriate nodes. The insertBefore and insertAfter methods always call disconnect on the node to insert before attempting to connect it to a linked list. This ensures the node is in a known state before assigning it to the Next or Prev property:
function disconnect(node) if ~isscalar(node)
2-36
error('Nodes must be scalar') end prevNode = node.Prev; nextNode = node.Next; if ~isempty(prevNode) prevNode.Next = nextNode; end if ~isempty(nextNode) nextNode.Prev = prevNode; end node.Next = []; node.Prev = []; end
For example, suppose you remove n2 from the three-node list discussed above (n1 n2 n3):
n2.disconnect;
Disconnect the nodes

disconnect removes n2 from the list and repairs the list with the following
steps:
n1 = n2.Prev; n3 = n2.Next; if n1 exists, then n1.Next = n3; if n3 exists, then n3.Prev = n1
2-37
Now the list is rejoined because n1 connects to n3 and n3 connects to n1. The final step is to ensure that n2.Next and n2.Prev are both empty (i.e., n2 is not connected):
% These properties have private SetAccess % so they can be set only within class methods n2.Next = []; n2.Prev = [];
Inserting Nodes
There are two methods for inserting nodes into the listinsertAfter and insertBefore. These methods perform similar operations, so this section describes only insertAfter in detail.
methods function insertAfter(newNode,nodeBefore) disconnect(newNode); newNode.Next = nodeBefore.Next; newNode.Prev = nodeBefore; if ~isempty(nodeBefore.Next) nodeBefore.Next.Prev = newNode; end nodeBefore.Next = newNode; end
How insertAfter Works. First insertAfter calls the disconnect method to ensure that the new node is not connected to any other nodes. Then, it assigns the newNode Next and Prev properties to the handles of the nodes that are after and before the newNode location in the list. For example, suppose you want to insert a new node, nnew, after an existing node, n1, in a list containing n1 n2. First, create nnew:
nnew = dlnode(rand(3));
Next, call insertAfter to insert nnew into the list after n1:
nnew.insertAfter(n1)
2-38
The insertAfter method performs the following steps to insert nnew in the list between n1 and n2:
% n1.Next is currently n2, set nnew.Next to n1.Next (which is n2) nnew.Next = n1.Next; % nnew.Prev must be set to n1 nnew.Prev = n1; % if n1.Next is not empty, then % n1.Next is still n2, so n1.Next.Prev is n2.Prev, which is set to nnew n1.Next.Prev = nnew; % n1.Next is now set to nnew n1.Next = nnew;
Newly inserted node
Displaying a Node on the Command Line

All objects call a default disp function, which displays information about the object on the command line (unless display is suppressed with a semicolon). The default disp function is not useful in this case because the Next and Prev properties contain other node objects. Therefore, the dlnode class overloads the default disp function by implementing its own disp class method. This disp method displays only a text message and the value of the Data property, when used with scalar objects, and array dimensions when used with object arrays.
function disp(node) % DISP Display a link node disp('Doubly-linked list node with data:') disp(node.Data) else % If node is an object array, display dimensions if (isscalar(node))
2-39
dims = size(node); ndims = length(dims); % Counting down in for loop avoids need to preallocate dimcell for k = ndims-1:-1:1 dimcell{k} = [num2str(dims(k)) 'x']; end dimstr = [dimcell{:} num2str(dims(ndims))]; disp([dimstr ' array of doubly-linked list nodes']); end end
Deleting a Node Object

MATLAB destroys a handle object when you reassign or delete its variable or when there are no longer any references to the object (see Handle Class Destructor on page 5-16 for more information). When you define a delete method for a handle class, MATLAB calls this method before destroying the object. The dlnode class defines a delete method because each dlnode object is a node in a doubly linked list. If a node object is going to be destroyed, the delete method must disconnect the node and repair the list before allowing MATLAB to destroy the node. The disconnect method already performs the necessary steps, so the delete method can simply call disconnect:
function delete(node) disconnect(node); end
Specializing the dlnode Class

The dlnode class implements a doubly linked list and provides a convenient starting point for creating more specialized types of linked lists. For example, suppose you want to create a list in which each node has a name. Rather than copying the code used to implement the dlnode class, and then expanding upon it, you can derive a new class from dlnode (i.e., subclass dlnode) to create a class that has all the features of dlnode and more. And because dlnode is a handle class, this new class is a handle class too.
2-40
NamedNode Class Definition

Open class definition file in the MATLAB editor. Use this link if you want to save and modify your version of the class. To use the class, create a folder named @NamedNode and save NamedNode.m to this folder. The parent folder of @NamedNode must be on the MATLAB path. Alternatively, save NamedNode.m to a path folder. The following class definition shows how to derive the NamedNode class from the dlnode class:
classdef NamedNode < dlnode properties Name = ''; % property to contain node name end methods function n = NamedNode (name,data) if nargin == 0 name = ''; data = []; end n = n@dlnode(data); % Initialize a dlnode object n.Name = name; end function disp(node) % Extend the dlnode disp method if (isscalar(node)) disp(['Node Name: ' node.Name]) disp@dlnode(node); % Call dlnode disp method else disp@dlnode(node); end end end % methods end % classdef % allow for the no argument case
The NamedNode class adds a Name property to store the node name and extends the disp method defined in the dlnode class. The constructor calls the class constructor for the dlnode class, and then assigns a value to the Name property. The NamedNode class defines default
2-41
values for the properties for cases when MATLAB calls the constructor with no arguments. See Basic Structure of Constructor Methods on page 7-23 for more information on defining class constructor methods.
Using NamedNode to Create a Doubly Linked List

Use the NamedNode class like the dlnode class, except you specify a name for each node object. For example:
n(1) = NamedNode('First Node',100); n(2) = NamedNode('Second Node',200); n(3) = NamedNode('Third Node',300);
Now use the insert methods inherited from dlnode to build the list:
n(2).insertAfter(n(1)) n(3).insertAfter(n(2))
A single node displays its name and data when you query its properties:
>> n(1).Next ans = Node Name: Second Node Doubly-linked list node with data: 200 >> n(1).Next.Next ans = Node Name: Third Node Doubly-linked list node with data: 300 >> n(3).Prev.Prev ans = Node Name: First Node Doubly-linked list node with data: 100
If you display an array of nodes, the NamedNode disp method displays only the dimensions of the array:
2-42
>> n n = 1x3 array of doubly-linked list nodes
2-43
Class for Graphing Functions

In this section... Commented Example Code on page 2-44 Class Definition Block on page 2-44 Using the topo Class on page 2-46 Behavior of the Handle Class on page 2-47 The class block is the code that starts with the classdef key word and terminates with the end key word. The following example illustrated a simple class definition that uses: Handle class Property set and get functions Use of a delete method for the handle object Static method syntax

You can display this class definition in a separate window that contains links to related sections in the documentations by clicking this link: Example with links Open class definition file in the MATLAB editor. Use this link if you want to save and modify your own version of the class.
Class Definition Block

The following code defines a class called topo. It is derived from handle so it is a handle class, which means it references the data it contains. See Using the topo Class on page 2-46 for information on how this class behaves.
classdef topo < handle % topo is a subclass of handle properties FigHandle % Store figure handle
2-44
FofXY % function handle Lm = [-2*pi 2*pi]; % Initial limits end % properties properties (Dependent, SetAccess = private) Data end % properties Dependent = true, SetAccess = private methods function obj = topo(fnc,limits) % Constructor assigns property values obj.FofXY = fnc; obj.Lm = limits; end % topo function set.Lm(obj,lim) % Lm property set function if else obj.Lm = lim; end end % set.Lm function data = get.Data(obj) % get function calculates Data % Use class name to call static method [x,y] = topo.grid(obj.Lm); matrix = obj.FofXY(x,y); data.X = x; data.Y = y; data.Matrix = matrix;% Return value of property end % get.Data function surflight(obj) % Graph function as surface obj.FigHandle = figure; surfc(obj.Data.X,obj.Data.Y,obj.Data.Matrix,... 'FaceColor',[.8 .8 0],'EdgeColor',[0 .2 0],... 'FaceLighting','phong'); ~(lim(1) < lim(2)) error('Limits must be monotonically increasing')
2-45
camlight left; material shiny; grid off colormap copper end % surflight method function delete(obj) % Delete the figure h = obj.FigHandle; if ishandle(h) delete(h); else return end end % delete end % methods methods (Static = true) % Define static method function [x,y] = grid(lim) inc = (lim(2)-lim(1))/35; [x,y] = meshgrid(lim(1):inc:lim(2)); end % grid end % methods Static = true end % topo class
Using the topo Class

See Commented Example Code on page 2-44 for information on using this class. This class is designed to display a combination surface/contour graph of mathematical functions of two variables evaluated on a rectangular domain of x and y. For example, any of the following functions can be evaluated over the specified domain (note that x and y have the same range of values in this example just for simplicity).
x.*exp(-x.^2 - y.^2); [-2 2] sin(x).*sin(y); [-2*pi 2*pi] sqrt(x.^2 + y.^2); [-2*pi 2*pi]
To create an instance of the class, passing a function handle and a vector of limits to the constructor. The easiest way to create a function handle for these functions is to use an anonymous function:
2-46
tobj = topo(@(x,y) x.*exp(-x.^2-y.^2),[-2 2]);
The class surflight method uses the object to create a graph of the function. The actual data required to create the graph is not stored. When the surflight method accesses the Data property, the propertys get function performs the evaluation and returns the data in the Data property structure fields. This data is then plotted. The advantage of not storing the data is the reduced size of the object.
Behavior of the Handle Class

The topo class is defined as a handle class. This means that instances of this class are handle objects that reference the underlying data store created by constructing the object. For example, suppose you create an instance of the class and create a copy of the object:
tobj = topo(@(x,y) x.*exp(-x.^2-y.^2),[-2 2]); a = tobj; surflight(a) % Call class method to create a graph
Now suppose you change the FofXY property so that it contains a function handle that points to another function:
2-47
tobj.FofXY = @(x,y) y.*exp(-x.^2-y.^2); % now multiply exp by y instead of x surflight(a)
Because a is a copy of the handle object tobj, changes to the data referenced by tobj also change the data referenced by a.
How a Value Class Differs

If topo were a value class, the objects tobj and a would not share data; each would have its own copy of the property values.
2-48
3
Class DefinitionSyntax Reference
Class Files on page 3-2 Class Components on page 3-5 Classdef Block on page 3-8 Properties on page 3-10 Methods and Functions on page 3-15 Events and Listeners on page 3-21 Specifying Attributes on page 3-23 Calling Superclass Methods on Subclass Objects on page 3-26 Representative Class Code on page 3-29 MATLAB Code Analyzer Warnings on page 3-31 Objects In Switch Statements on page 3-34 Using the Editor and Debugger with Classes on page 3-42 Modifying and Reloading Classes on page 3-43 Compatibility with Previous Versions on page 3-46 Comparing MATLAB with Other OO Languages on page 3-50
Class Files
In this section... Options for Class Folders on page 3-2 Grouping Classes with Package Folders on page 3-3 More Information on Class Folders on page 3-4
Options for Class Folders

There are two basic ways to specify classes with respect to folders: Creating a single, self-contained class definition file in a folder on the MATLAB path. Distributing a class definition to multiple files in an @ folder inside a path folder.
Creating a Single, Self-Contained Class Definition File

Create a single, self-contained class definition file in a folder on the MATLAB path. The name of the file must match the class (and constructor) name and must have the .m extension. Define the class entirely in this file. You can put other single-file classes in this folder. The following diagram shows an example of this folder organization.
pathfolder is a folder on the MATLAB path.
pathfolder ClassNameA.m ClassNameB.m ClassNameC.m ... ordinaryFunction.m Contains classdef and methods for ClassNameA Contains classdef and methods for ClassNameB Contains classdef and methods for ClassNameC A function on the path
See Methods in Separate Files on page 7-8 for more information on using multiple files to define classes.
3-2
Class Files
Distributing the Class Definition to Multiple Files

If you use multiple files to define a class, put all the class-definition files (the file containing the classdef and all class method files) in a single @ClassName folder. That @-folder must be inside a folder that is on the MATLAB path. You can define only one class in an @-folder.
pathfolder @ClassNameA ClassNameA.m classMethod.m ClassNameB.m Contains classdef Class method in separate file Contains entire class definition
A path folder can contain classes defined in both @-folders and single files without an @-folder.
Grouping Classes with Package Folders

The parent folder to a package folder is on the MATLAB path, but the package folder is not. Package folders (which always begin with a + character) can contain multiple class definitions, package-scoped functions, and other packages. A package folder defines a new name space in which you can reuse class names. Use the package name to refer to classes and functions defined in package folders (for example, packagefld1.ClassNameA(), packagefld2.packageFunction()).
3-3
pathfolder +packagefld1 @ClassNameA ClassNameA.m classMethod.m ClassNameB.m +packagefld2 packageFunction.m ClassNameA.m ClassNameB.m Contains classdef Class method in separate file Contains entire class definition Defines a new name space
More Information on Class Folders

See Organizing Classes in Folders on page 4-15 for more information on class folders and see Packages Create Namespaces on page 4-20 for information on using classes contained in package folders. See Methods In Separate Files on page 3-16 for the syntax used to define methods external to the classdef file.
3-4
Class Components
Class Components
In this section... Class Building Blocks Defining Class Members on page 3-5 More In Depth Information on page 3-6
Class Building Blocks Defining Class Members

The basic components in the class definition are blocks describing the whole class and specific aspects of its definition: classdef block contains the class definition within a file that starts with the classdef keyword and terminates with the end keyword. See Classdef Block on page 3-8 for more syntax information.
classdef (ClassAttributes ClassName) ClassName ... end
properties block (one for each unique set of attribute specifications) contains property definitions, including optional initial values. The properties block starts with the properties keyword and terminates with the end keyword. See Properties on page 3-10 for more syntax information.
classdef ClassName properties (PropertyAttributes) ... end ... end
methods block (one for each unique set of attribute specifications) contains function definitions for the class methods. The methods block starts with the methods keyword and terminates with the end keyword. See The Methods Block on page 3-15 for more syntax information.
classdef ClassName methods (MethodAttributes) ...
3-5
end ... end
events block (one for each unique set of attribute specifications) contains the names of events that this class declares. The events block starts with the events keyword and terminates with the end keyword. See Specifying Events on page 3-21 for more syntax information.
classdef ClassName events (EventAttributes) ... end ... end
enumeration block contains the enumeration members defined by the class. The enumeration block starts with the enumeration keyword and terminates with the end keyword. See Using Enumeration Classes on page 12-5 for more information.
classdef Boolean < logical enumeration No (0) Yes (1) end end properties, methods, events, and enumeration are keywords only within a classdef block.
More In Depth Information

Class Definition on page 4-4 for more detail on class syntax. Defining Properties on page 6-5 for information on specifying properties. How to Use Methods on page 7-2 for information on specifying methods.
3-6
Class Components
Events and Listeners Syntax and Techniques on page 9-18 for information on the use of events. Defining Named Values on page 12-2 for information on creating and using enumeration classes. Attribute Tables for a list of all attributes.
3-7
Classdef Block
In this section... Specifying Attributes and Superclasses on page 3-8 Assigning Class Attributes on page 3-8 Specifying Superclasses on page 3-9
Specifying Attributes and Superclasses

The classdef block contains the class definition. The classdef line is where you specify: Class attributes Superclasses The classdef block contains the properties, methods, and events subblocks.
Assigning Class Attributes

Class attributes modify class behavior in some way. Assign values to class attributes only when you want to change their default value. No change to default attribute values:
classdef class_name ... end
One or more attribute values assigned:

classdef (attribute1 = value,...) ... end
See Class Attributes on page 4-6 for a list of attributes and a discussion of the behaviors they control.
3-8
Classdef Block
Specifying Superclasses
To define a class in terms of one or more other classes by specifying the superclasses on the classdef line:
classdef class_name < superclass_name ... end
See Creating Subclasses Syntax and Techniques on page 10-7 for more information.
3-9
Properties
In this section... What You Can Define on page 3-10 Initializing Property Values on page 3-10 Defining Default Values on page 3-11 Assigning Property Values from the Constructor on page 3-11 Initializing Properties to Unique Values on page 3-12 Property Attributes on page 3-12 Property Access Methods on page 3-13 Referencing Object Properties Using Variables on page 3-13
What You Can Define

You can control aspects of property definitions in the following ways: Specifying a default value for each property individually Assigning attribute values on a per block basis Defining methods that execute when the property is set or queried Note Always use case sensitive property names in your MATLAB code.
Initializing Property Values

There are two basic approaches to initializing property values: In the property definition MATLAB evaluates the expression only once and assigns the same value to the property of every instance. See Defining Default Values on page 3-11. In the class constructor MATLAB evaluates the assignment expression for each instance, which ensures that each instance has a unique value. See Assigning Property Values from the Constructor on page 3-11.
3-10
Properties
Defining Default Values

Within a properties block, you can control an individual propertys default value. Default values can be constant values or MATLAB expressions. Expressions cannot reference variables. For example:
classdef class_name properties PropertyName % No default value assigned PropertyName = 'some text'; PropertyName = sin(pi/12); % Expression returns default value end end
Note Evaluation of property default values occurs only when the value is first needed, and only once when MATLAB first initializes the class. MATLAB does not reevaluate the expression each time you create a class instance. See When MATLAB Evaluates Expressions on page 4-11 for more information on the evaluation of expressions that you assign as property default values. MATLAB sets property values not specified in the class definition to empty ([]).
Assigning Property Values from the Constructor

To assign values to a property from within the class constructor, reference the object that the constructor returns (the output variable obj):
classdef MyClass properties PropertyOne end methods function obj = MyClass(intval) obj.PropertyOne = intval; end end
3-11
end
When you assign an object property from the class constructor, MATLAB evaluates the assignment statement for each instance created. Assign property values in the constructor if you want each object to contain a unique instance of a handle object. See Referencing the Object in a Constructor on page 7-18 for more information on constructor methods.
Initializing Properties to Unique Values

MATLAB assigns properties to the specified default values only once when MATLAB loads the class definition. Therefore, if you initialize a property value with a handle-class constructor, MATLAB calls this constructor only once and every instance references the same handle object. If you want a property value to be initialized to a new instance of a handle object each time you create an object, assign the property value in the constructor.
Property Attributes
All properties have attributes that modify certain aspects of the propertys behavior. Specified attributes apply to all properties in a particular properties block. For example:
classdef class_name properties
PropertyName % No default value assigned PropertyName = sin(pi/12); % Expression returns default value
end properties (SetAccess = private, GetAccess = private) Stress Strain end end
In this case, only methods in the same class definition can modify and query the Stress and Strain properties. This restriction exists because the class defines these properties in a properties block with SetAccess and GetAccess attributes set to private.
3-12
Properties
Table of Property Attributes on page 6-8 provides a description of property attributes.
Property Access Methods

You can define methods that MATLAB calls whenever setting or querying a property value. Define property set access or get access methods in methods blocks that specify no attributes and have the following syntax:
methods function value = get.PropertyName(object) ... end function obj = set.PropertyName(obj,value) ... end end
MATLAB does not call the property set access method when assigning the default value specified in the propertys definition block. If a handle class defines the property, the set access method does not need to return the modified object. Property Setter and Getter Methods on page 6-14 for more information on these methods. Defining Properties on page 6-5 for information on properties.
Referencing Object Properties Using Variables

MATLAB can resolve a property name from a char variable using an expression of the form:
object.(PropertyNameVar)
where PropertyNameVar is a variable containing the name of a valid object property. Use this syntax when passing property names as arguments:
PropName = 'KeyType';
3-13
function o = getPropValue(obj,PropName) ... o = obj.(PropName); % Returns value of KeyType property ... end
3-14
Methods and Functions

In this section... The Methods Block on page 3-15 Method Calling Syntax on page 3-16 Methods In Separate Files on page 3-16 Private Methods on page 3-18 More Detailed Information On Methods on page 3-18 Class-Related Functions on page 3-19 Overloading Functions and Operators on page 3-19
The Methods Block

Define methods as MATLAB functions within a methods block, inside the classdef block. The constructor method has the same name as the class and returns an object. You can assign values to properties in the class constructor. Terminate all method functions with an end statement.
classdef ClassName methods function obj = ClassName(arg1,arg2,...) obj.Prop1 = arg1; ... end function normal_method(obj,arg1,...) ... end end methods (Static = true) function static_method(arg1,...) ... end end end
3-15
Method Calling Syntax

MATLAB differs from languages like C++ and Java in that there is no special hidden class instance passed to all methods. You must pass an object of the class explicitly to the method. The left most argument does not need to be the class instance, and the argument list can have multiple objects. See Determining Which Method Is Invoked on page 7-9 for more information. See alsoStatic Methods on page 7-25 for information on methods that do not require instances of their class. Note Always use case sensitive method names in your MATLAB code.
Methods In Separate Files

You can define class methods in files that are separate from the class definition file, with certain exceptions (see Methods That Must Be In the classdef File on page 3-17). To use multiple files for a class definition, put the class files in a folder having a name beginning with the @ character followed by the name of the class. Ensure that the parent folder of the @-folder is on the MATLAB path, or, if the @-folder is contained in one or more package folders, then the top-level package folder must be on the MATLAB path. For example, the folder @MyClass must contain the file MyClass.m (which contains the classdef block) and can contain other methods and function defined in files having a .m extension. For example, the folder @MyClass might contain a number of files:
@MyClass/MyClass.m @MyClass/subsref.m @MyClass/subsasgn.m @MyClass/horzcat.m @MyClass/vertcat.m @MyClass/myFunc.m
3-16
Define the Method Like Any Function

To define a method in a separate file in the class @-folder, create the function in a separate file, but do not use a method block in that file. Name the file with the function name, as with any function.
Methods That Must Be In the classdef File

You must put the following methods in the classdef file, not in separate files: Class constructor Delete method All functions that use dots in their names, including:
Converter methods that convert to classes contained in packages, which must use the package name as part of the class name. Property set and get access methods (Property Setter and Getter Methods on page 6-14)
Specifying Method Attributes in classdef File

If you specify method attributes for a method that you define in a separate file, include the method signature in a methods block in the classdef block. For example, the following code shows a method with Access set to private in the methods block. The method implementation resides in a separate file. Do not include the function or end keywords in the methods block, just the function signature showing input and output arguments.
classdef ClassName % In a methods block, set the method attributes % and add the function signature methods (Access = private) output = myFunc(obj,arg1,arg2) end end
In a file named myFunc.m, in the @ClassName folder, define the function:

function output = myFunc(obj,arg1,arg2) ... end
3-17
Include the method signature in the file with the classdef block only if you want to specify attributes for that method. Otherwise, you can implement the method as a function in a separate file in the @-folder.
Defining Static Methods in Separate Files

To create a static method, set the functions Static attribute to true. List any static methods that you define in separate files in the @-class folder. List these methods in the static methods block in the classdef file. Include the input and output arguments with the function name. For example:
classdef ClassName ... methods (Static) output = staticFunc1(arg1,arg2) staticFunc2 end
You would then define the functions in separate files using the same function signature. For example:
function output = staticFunc1(arg1,arg2) ... end
Using Separate Files for Methods

The example, Update Graphs Using Events and Listeners on page 9-35 uses multiple files for class definition.
Private Methods
Use the Access method attribute to create a private method. You do not need to use a private folder. See Method Attributes on page 7-5 for a list of method attributes.
More Detailed Information On Methods

See How to Use Methods on page 7-2 for more information about methods.
3-18
Class-Related Functions
You can define functions that are not class methods in the file that contains the class definition (classdef). Define local functions outside of the classdef - end block, but in the same file as the class definition. Functions defined in classdef files work like local functions. You can call these functions from anywhere in the same file, but they are not visible outside of the file in which you define them. Local functions in classdef files are useful for utility functions that you use only within that file. These functions can take or return arguments that are instances of the class but, it is not necessary, as in the case of ordinary methods. For example, the following code defines myUtilityFcn outside the classdef block:
classdef MyClass properties PropName end methods function obj = MyClass(arg1) obj.PropName = arg1; end end % methods end % classdef function myUtilityFcn ... end
You also can create package functions, which require you to use the package name when calling these functions. See Packages Create Namespaces on page 4-20 for more information on packages
Overloading Functions and Operators

Overload MATLAB functions for your class by defining a class method with the same name as the function you want to overload. MATLAB dispatches to the class method when the function is called with an instance of the class. See Overloading Functions for Your Class on page 7-27 for more information.
3-19
You can also overload MATLAB arithmetic, logical, relational, and indexing operators by defining class methods with the appropriate names. See Implementing Operators for Your Class on page 15-31 for a list of the functions to overload. See the handle class for a list of operations defined for that class, which are inherited by all classes deriving from handle.
3-20
Events and Listeners
Events and Listeners

In this section... Specifying Events on page 3-21 Listening for Events on page 3-21
Specifying Events
To define an event, you declare a name for the event in the events block. Then one of the class methods triggers the event using the notify method, which is method inherited from the handle class. Only classes derived from the handle class can define events. For example, the following class: Defines an event named StateChange Triggers the event using the inherited notify method.
classdef class_name < handle % Subclass handle events % Define an event called StateChange StateChange end ... methods function upDateGUI(obj) ... % Broadcast notice that StateChange event has occurred notify(obj,'StateChange'); end end end
Listening for Events

Any number of objects can be listening for the StateChange event to occur. When notify executes, MATLAB calls all registered listener callbacks and passes the handle of the object generating the event and an event structure to
3-21
these functions. To register a listener callback, use the addlistener method of the handle class.
addlistener(event_obj,'StateChange',@myCallback)
See Learning to Use Events and Listeners on page 9-2 for more information on using events and listeners.
3-22
Specifying Attributes
In this section... Attribute Syntax on page 3-23 Attribute Descriptions on page 3-23 Attribute Values on page 3-24 Simpler Syntax for true/false Attributes on page 3-24
Attribute Syntax
For a quick reference to all attributes, see Attribute Tables. Attributes modify the behavior of classes and class components (properties, methods, and events). Attributes enable you to define useful behaviors without writing complicated code. For example, you can create a read-only property by setting its SetAccess attribute to private, but leaving its GetAccess attribute set to public (the default):
properties (SetAccess = private) ScreenSize = getScreenSize; end
All class definition blocks (classdef, properties, methods, and events) support specific attributes and all attributes have default values. Specify attribute values only in cases where you want to change from the default value to another predefined value. Note Specify the value of a particular attribute only once in any component block.
Attribute Descriptions
For lists of supported attributes, see: Class Attributes on page 4-6 Property Attributes on page 6-8
3-23
Method Attributes on page 7-5 Event Attributes on page 9-16
Attribute Values
When you specify attribute values, these values affect all the components defined within the definition block. For example, the following property definition blocks set the: AccountBalance property SetObservable attribute to true SSNumber and CreditCardNumber properties Hidden attribute to true and SetAccess attribute to private. Defining properties with different attribute settings requires multiple
properties blocks.
properties (SetObservable = true) AccountBalance end properties (SetAccess = private, Hidden = true) SSNumber CreditCardNumber end
Specified multiple attributes in a comma-separated list, as shown in the previous example. When specifying class attributes, place the attribute list directly after the classdef keyword:
classdef (Sealed = true) myclass ... end
Simpler Syntax for true/false Attributes

You can use a simpler syntax for attributes whose values are true or false the attribute name alone implies true and adding the not operator (~) to the name implies false. For example:
methods (Static)
3-24
... end
is the same as:

methods (Static = true) ... end
Use the not operator before an attribute name to define it as false:

methods (~Static) ... end
is the same as:

methods (Static = false) ... end
All attributes that take a logical value (that is, true or false) have a default value of false. Therefore, specify an attribute only if you want to set it to true.
3-25
Calling Superclass Methods on Subclass Objects

In this section... Calling a Superclass Constructor on page 3-26 Calling Superclass Methods on page 3-27
Calling a Superclass Constructor

If you create a subclass object, MATLAB calls the superclass constructor to initialize the superclass part of the subclass object. By default, MATLAB calls the superclass constructor without arguments. If you want the superclass constructor called with specific arguments, explicitly call the superclass constructor from the subclass constructor. The call to the superclass constructor must come before any other references to the object. The syntax for calling the superclass constructor uses an @ symbol:
classdef MySub < MySuperClass methods function obj = MySub(arg1,arg2,...) obj = obj@MySuperClass(SuperClassArguments); ... end % MySub end % methods end % classdef
Interpret this syntax as meaning, the MySub object arrives at the MySuperClass constructor , which constructs the MySuperClass part of the object using the specified arguments.
3-26
Calling Superclass Methods on Subclass Objects
obj = obj@MySuperClass(SuperClassArguments);
Object returned from superclass
Name of superclass Object being constructed Superclass constructor arugment list
See Constructing Subclasses on page 7-20 for more information.
Calling Superclass Methods

You can call a superclass method from a subclass method if both methods have the same name. From the subclass, reference the method name and superclass name with the @ symbol. See Modifying Superclass Methods on page 10-14 for more information on when to call superclass methods. For example, a subclass can call a superclass disp method to implement the display of the superclass part of the object, and then add code to display the subclass part:
classdef MySub < MySuperClass methods function disp(obj) disp@MySuperClass(obj) ... end % disp end % methods end % classdef
This diagram illustrates how to call the superMethod defined at MySuperClass.
3-27
superMethod@MySuperClass(obj)
Superclass name Superclass method Object passed to the superclass method
3-28
Representative Class Code
Representative Class Code

Example of Class Definition Syntax
The following code shows the syntax of a typical class definition. This example is not a functioning class because it references functions that it does not implement. The purpose of this section is to illustrate various syntactic constructions.
classdef (ConstructOnLoad) Employee < handle % Class help goes here properties Name % Property help goes here end properties (Dependent) JobTitle end properties (Transient) OfficeNumber end properties (SetAccess = protected, GetAccess = private) EmpNumber end events BackgroundAlert end methods function Eobj = Employee(name) % Method help here Eobj.Name = name; Eobj.EmpNumber = employee.getEmpNumber; end function result = backgroundCheck(obj) result = queryGovDB(obj.Name,obj.SSNumber);
3-29
if result == false notify(obj,'BackgroundAlert'); end end function jobt = get.JobTitle(obj) jobt = currentJT(obj.EmpNumber); end function set.OfficeNumber(obj,setvalue) if isInUse(setvalue) error('Not available') else obj.OfficeNumber = setvalue; end end end methods (Static) function num = getEmpNumber num = queryDB('LastEmpNumber') + 1; end end end
3-30
MATLAB Code Analyzer Warnings

In this section... Syntax Warnings and Property Names on page 3-31 Warnings Caused by Variable/Property Name Conflicts on page 3-31 Exception to Variable/Property Name Rule on page 3-32
Syntax Warnings and Property Names

The MATLAB Code Analyzer helps you optimize your code and avoid syntax errors while you write code. It is useful to understand some of the rules that the Code Analyzer applies in its analysis of class definition code. This understanding helps you avoid situations in which MATLAB allows code that is undesirable.
Warnings Caused by Variable/Property Name Conflicts

The Code Analyzer warns about the use of variable names in methods that match the names of properties. For example, suppose a class defines a property called EmployeeName and in this class, there is a method that uses EmployeeName as a variable:
properties EmployeeName end methods function someMethod(obj,n) EmployeeName = n; end end
While the previous function is legal MATLAB code, it results in Code Analyzer warnings for two reasons: The value of EmployeeName is never used EmployeeName is the name of a property that is used as a variable
3-31
If the function someMethod contained the following statement instead:

obj.EmployeeName = n;
The Code Analyzer generates no warnings. If you change someMethod to:

function EN = someMethod(obj) EN = EmployeeName; end
The Code Analyzer returns only one warning, suggesting that you might actually want to refer to the EmployeeName property. While this version of someMethod is legal MATLAB code, it is confusing to give a property the same name as a function. Therefore, the Code Analyzer provides a warning suggesting that you might have intended the statement to be:
EN = obj.EmployeeName;
Exception to Variable/Property Name Rule

Suppose you define a method that returns a value of a property and uses the name of the property for the output variable name. For example:
function EmployeeName = someMethod(obj) EmployeeName = obj.EmployeeName; end
The Code Analyzer does not warn when a variable name is the same as a property name when the variable is: An input or output variable A global or persistent variable In these particular cases, the Code Analyzer does not warn you that you are using a variable name that is also a property name. Therefore, a coding error like the following:
3-32
function EmployeeName = someMethod(obj) EmployeeName = EmployeeName; % Forgot to include obj. end
does not trigger a warning from the Code Analyzer.
3-33
Objects In Switch Statements

In this section... Evaluating the Switch Statement on page 3-34 Defining the eq Method on page 3-36 Enumerations in Switch Statements on page 3-38 Functions to Test Objects on page 3-40 Functions to Query Class Members on page 3-41
Evaluating the Switch Statement

MATLAB enables you to use objects in switch and case statements if the objects class defines an eq method. The eq method implements the == operation on objects of that class. For objects, switch_expression == case_expression defines how MATLAB evaluates switch and cases statements. Note You do not need to define eq methods for enumeration classes. See Enumerations in Switch Statements on page 3-38.
Handle Objects in Switch Statements

All classes derived from the handle class inherit an eq method. The expression,
h1 == h2
is true if h1 and h2 are handles for the same object. For example, the BasicHandle class derives from handle:
classdef BasicHandle < handle properties Prop1 end
3-34
methods function obj = BasicHandle(val) if nargin > 0 obj.Prop1 = val; end end end end
Create a BasicHandle object and use it in a switch statement:

h1 = BasicHandle('Handle Object'); h2 = h1; switch h1 case h2 disp('h2 is selected') otherwise disp('h2 not selected') end
The result is:

h2 is selected
Object Must Be Scalar

The switch statements work only with scalar objects. For example:
h1(1) = BasicHandle('Handle Object'); h1(2) = BasicHandle('Handle Object'); h1(3) = BasicHandle('Handle Object'); h2 = h1; switch h1 case h2 disp('h2 is selected') otherwise disp('h2 not selected') end SWITCH expression must be a scalar or string constant.
3-35
In this case, h1 is not scalar. Use isscalar to determine if an object is scalar before entering a switch statement.
Defining the eq Method

To enable the use of value-class objects in switch statements, implement an eq method for the class. Use the eq method to determine what constitutes equality of two object of the class.
Behave Like a Built-in Type

Some MATLAB functions also use the built-in == operator in their implementation. Therefore, your implementation of eq should be replaceable with the built-in eq to enable objects of your class work like built-in types in MATLAB code.
Design of eq
Implement the eq method to returns a logical array representing the result of the == comparison. For example, the SwitchOnVer class implements an eq method that returns true for the == operation if the value of the Version property is the same for both objects. In addition, eq works with arrays the same way as the built-in eq. For the following expression:
obj1 == obj2
The eq method works like this: If both obj1 and obj2 are scalar, eq returns a scalar value. If both obj1 and obj2 are nonscalar arrays, then these arrays must have the same dimensions, and eq returns an array of the same size. If one input argument is scalar and the other is a nonscalar array, then eq treats the scalar object as if it is an array having the same dimensions as the nonscalar array.
3-36
Implementation of eq
Here is a class that implements an eq method. Ensure your implementation contains appropriate error checking for the intended use.
classdef SwitchOnVer properties Version end methods function obj = SwitchOnVer(ver) if nargin > 0 obj.Version = ver; end end function bol = eq(obj1,obj2) if ~strcmp(class(obj1),class(obj2)) error('Objects are not of the same class') end s1 = numel(obj1); s2 = numel(obj2); if s1 == s2 bol = false(size(obj1)); for k=1:s1 if obj1(k).Version == obj2(k).Version bol(k) = true; else bol(k) = false; end end elseif s1 == 1 bol = scalarExpEq(obj2,obj1); elseif s2 == 1 bol = scalarExpEq(obj1,obj2); else error('Dimension missmatch') end function ret = scalarExpEq(ns,s) % ns is nonscalar array % s is scalar array ret = false(size(ns));
3-37
n = numel(ns); for kk=1:n if ns(kk).Version == s.Version ret(kk) = true; else ret(kk) = false; end end end end end end
Use SwitchOnVer objects in switch statements:

% Create known versions of objects ov1 = SwitchOnVer(1.0); ov2 = SwitchOnVer(2.0); ov3 = SwitchOnVer(3.0); ... ... if isscalar(objIn) switch(objIn) case ov1 disp('This is version 1.0') case ov2 disp('This is version 2.0') case ov3 disp('This is version 3.0') otherwise disp('There is no version') end else error('Input object must be scalar') end
Enumerations in Switch Statements

MATLAB enables you to use enumerations in switch statements without requiring an explicitly defined eq method for the enumeration class.
3-38
For example, the WeeklyPlanner class defines enumerations for five days of the week. The switch/case statements in the todaySchedule static method dispatch on the enumeration member corresponding to the current day of the week. The date and datestr functions return a character string with the name of the current day.
classdef WeeklyPlanner enumeration Monday, Tuesday, Wednesday, Thursday, Friday end methods (Static) function todaySchedule dayName = datestr(date,'dddd'); dayEnum = WeeklyPlanner.(dayName); switch dayEnum case WeeklyPlanner.Monday disp('Monday schedule') case WeeklyPlanner.Tuesday disp('Tuesday schedule') case WeeklyPlanner.Wednesday disp('Wednesday schedule') case WeeklyPlanner.Thursday disp('Thursday schedule') case WeeklyPlanner.Friday disp('Friday schedule') end end end end
Call todaySchedule to display todays schedule:

WeeklyPlanner.todaySchedule
Enumerations Derived from Built-In Types

Enumeration classes that derived from built-in types inherit the superclass eq method. For example, the FlowRate class derives from int32:
classdef FlowRate < int32
3-39
enumeration Low (10) Medium (50) High (100) end end
The switchEnum function switches on the input argument, which can be a FlowRate enumeration value.
function switchEnum(inpt) switch inpt case 10 disp('Flow = 10 cfm') case 50 disp('Flow = 50 cfm') case 100 disp('Flow = 100 cfm') end end
Call switchEnum with an enumerated value:

switchEnum(FlowRate.Medium) Flow = 50 cfm
Functions to Test Objects

These functions provide logical tests, which are useful when using objects in ordinary functions. Function
isa isequal
Description Determine whether an argument is an object of specific class. Determine if two objects are equal, which means both objects are of the same class and size and their corresponding property values are equal. Determine whether input is a MATLAB object
isobject
3-40
Functions to Query Class Members

These functions provide information about object class members. Function
class enumeration events methods methodsview properties
Description Return class of object. Display class enumeration members and names. List event names defined by the class. List methods implemented by the class. List methods in separate window. List class property names.
3-41
Using the Editor and Debugger with Classes

Referring to Class Files
Define classes in files just like scripts and functions. To use the editor or debugger with a class file, use the full class name. For example, suppose the file for a class, myclass.m is in the following location:
+PackFld1/+PackFld2/@myclass/myclass.m
To open myclass.m in the MATLAB editor, you could reference the file using dot-separated package names:
edit PackFld1.PackFld2.myclass
You could also use path notation:

edit +PackFld1/+PackFld2/@myclass/myclass
If myclass.m is not in an @-folder, then enter:

edit +PackFld1/+PackFld2/myclass
To refer to functions inside a package folder, use dot or path separators:

edit PackFld1.PackFld2.packFunction edit +PackFld1/+PackFld2/packFunction
To refer to a function defined in its own file inside of a class @-folder, use:
edit +PackFld1/+PackFld2/@myclass/myMethod
Debugging Class Files

For debugging, dbstop accepts any of the file specifications used by the edit command. See Modifying and Reloading Classes on page 3-43 for information about clearing class.
3-42
Modifying and Reloading Classes

Ensuring MATLAB Uses Your Changes
There is only one class definition for a given class in MATLAB at any given time. When you create an instance of a class, MATLAB loads the class definition. So as long as instances of that class exist, MATLAB does not reload the class definition.
Clear Class Instances

When you modify a class definition, the current MATLAB session continues to use the original class definition until you clear all objects of that class. For example, if obj1 and obj2 are instances of a class for which you have modified the class definition, clear those objects so MATLAB can use your changes. Use the clear command to remove only those instances:
clear obj1 obj2
Modifying a class definition includes doing any of the following: Changing class member attributes Adding, deleting, or changing the names of properties, methods, or events Changing class inheritance Changing the definition of a superclass (requires you to clear subclass objects) If there are no class instances, MATLAB applies changes in the code immediately. If there are instances, you must clear those objects before MATLAB applies your changes. See Metaclass Object Lifecycle on page 14-4 for related information.
Clear Classes
When you issue the clear classes command, MATLAB clears: The current workspace of all variables
3-43
All functions, which can have persistent variables holding class instances (unless the function is locked) All classes that are not instantiated However, it is possible that your MATLAB session is holding instances of the class that the clear classes command does not clear. For example, suppose you change the definition of MyClass after saving an instance of this class in a Handle Graphics objects UserData property:
obj = MyClass; % User-defined class that you are editing h = uicontrol('Style','pushbutton'); set(h,'UserData',obj) clear classes Warning: Objects of 'MyClass' class exist. Cannot clear this class or any of its super-classes.
MATLAB issues a warning stating that it cannot apply your changes because it cannot clear the class. Clear the instance of MyClass before calling clear classes. For example, you can use the close all command to remove the object or reset the UserData property to another value:
% First, get the handle of the uicontrol, which was cleared h = findobj('Type','uicontrol','Style','pushbutton'); set(h,'UserData',[])
Now you can issue the clear classes command.
Places That Can Hold Instances

You can remove class instances from your workspace using the clear obj... command. However, as the preceding example shows, objects can be held in various ways. Clear all instances before MATLAB applies your new class definition. Here are some suggestions for finding and clearing class instances: Persistent Variables. Persistent variables can hold objects. Clear persistent variables using clear functions. If the function containing the persistent variable is locked, then unlock the function (using munlock) before clearing it.
3-44
Locked Functions. Functions can contain objects in their workspace. If the function is locked (with mlock), unlock it (using munlock) so that MATLAB can clear the instance. Use clear functions once you have unlocked the function. Default Property Values. When you specify a default value in a properties definition block, MATLAB evaluates the expression that defines the default value once when loading the class. Clear this value using the clear classes command. Constant Properties. When you define a constant property (property Constant attribute set to true) whose value is an object, MATLAB creates the instance when loading the class. Clear this instance using the clear classes command. Handle Graphics Objects. Handle Graphics objects can contain class instances in UserData properties, in Application Data, or created in callback functions. Issuing the close all command removes the Handle Graphics objects, unless these objects enable hidden handles. See the close command for more information. You can remove Application Data using the rmappdata function. Simulink Models. Models can contain class instances. Use close_system to close the model so that MATLAB can apply your changes.
3-45
Compatibility with Previous Versions

In this section... New Class-Definition Syntax Introduced with MATLAB Software Version 7.6 on page 3-46 Changes to Class Constructors on page 3-47 New Features Introduced with Version 7.6 on page 3-48 Examples of Old and New on page 3-48
New Class-Definition Syntax Introduced with MATLAB Software Version 7.6

MATLAB software Version 7.6 introduces a new syntax for defining classes. This new syntax includes: The classdef keyword begins a block of class-definitions code. An end statement terminates the class definition. Within the classdef code block, properties, methods, and events are also keywords delineating where you define the respective class members.
Cannot Mix Class Hierarchy

It is not possible to create class hierarchies that mix classes defined before Version 7.6 and current class definitions that use classdef. Therefore, you cannot subclass an old class to create a version of the new class.
Only One @-Folder per Class

For classes defined using the new classdef keyword, an @-folder shadows all @-folders that occur after it on the MATLAB path. Classes defined in @-folders must locate all class files in that single folder. However, classes defined in @-folders continue to take precedence over functions and scripts having the same name, even those function and scripts that come before them on the path.
3-46
Private Methods
You do not need to define private folders in class folders in Version 7.6. You can set the methods Access attribute to private instead.
Changes to Class Constructors

Class constructor methods have two major differences. Class constructors: Do not use the class function. Must call the superclass constructor only if you want to pass arguments to its constructor. Otherwise, no call to the superclass constructor is necessary.
Example of Old and New Syntax

Compare the following two Stock constructor methods. The Stock class is a subclass of the Asset class, which requires arguments passed to its constructor. Constructor Function Before Version 7.6
function s = Stock(description,num_shares,share_price) s.NumShares = num_shares; s.SharePrice = share_price; % Construct Asset object a = Asset(description,'stock',share_price*num_shares); % Use the class function to define the stock object s = class(s,'Stock',a);
Write the same Stock class constructor as shown here. Define the inheritance on the classdef line and define the constructor within a methods block. Constructor Function for Version 7.6
classdef Stock < Asset ... methods function s = Stock(description,num_shares,share_price) % Call superclass constructor to pass arguments
3-47
s = s@Asset(description,'stock',share_price*num_shares); s.NumShares = num_shares; s.SharePrice = share_price; end % End of function end % End of methods block end % End of classdef block
New Features Introduced with Version 7.6

Properties: How to Use Properties on page 6-2 Handle classes: Comparing Handle and Value Classes on page 5-2 Events and listeners: Events and Listeners Concepts on page 9-11 Class member attributes: Attribute Tables Abstract classes: Defining Abstract Classes on page 10-77 Dynamic properties: Dynamic Properties Adding Properties to an Instance on page 6-27 Ability to subclass MATLAB built-in classes: Creating Subclasses Syntax and Techniques on page 10-7 Packages for scoping functions and classes: Packages Create Namespaces on page 4-20. MATLAB does not support packages for classes created before MATLAB Version 7.6 (that is, classes that do not use classdef). The JIT/Accelerator supports objects defined only by classes using classdef.
Examples of Old and New

The MATLAB Version 7.6 implementation of classes uses different syntax from previous releases. However, classes written in previous versions continue to work. Most of the code you use to implement the methods is likely to remain the same, except where you take advantage of new features. The following sections reimplement examples using the latest syntax. The original MATLAB Classes and Objects documentation implemented these same examples and provide a comparison of old and new syntax.
3-48
A Polynomial Class on page 17-2 A Simple Class Hierarchy on page 18-2 Containing Assets in a Portfolio on page 18-19
3-49
Comparing MATLAB with Other OO Languages

In this section... Some Differences from C++ and Java Code on page 3-50 Modifying Objects on page 3-51 Common Object-Oriented Techniques on page 3-56
Some Differences from C++ and Java Code

The MATLAB programming language differs from other object-oriented languages, such as C++ or Java in some important ways.
Public Properties
Unlike fields in C++ or the Java language, you can use MATLAB properties to define a public interface separate from the implementation of data storage. You can provide public access to properties because you can define set and get access methods that execute automatically when assigning or querying property values. For example, the following statement:
myobj.Material = 'plastic';
assigns the string plastic to the Material property of myobj. Before making the actual assignment, myobj executes a method called set.Material (assuming the class of myobj defines this method), which can perform any necessary operations. See Property Access Methods on page 6-14 for more information on property access methods. You can also control access to properties by setting attributes, which enable public, protected , or private access. See Property Attributes on page 6-8 for a full list of property attributes.
No Implicit Parameters
In some languages, one object parameter to a method is always implicit. In MATLAB, objects are explicit parameters to the methods that act on them.
3-50
Dispatching
In MATLAB classes, method dispatching is not based on method signature, as it is in C++ and Java code. When the argument list contains objects of equal precedence, MATLAB software uses the left-most object to select the method to call. However, if the class of that argument is superior to the other arguments, MATLAB can dispatch to a method of an argument in any position within an argument list. See Class Precedence on page 4-18 for more information.
Calling Superclass Method

In C++, you call a superclass method using the scoping operator:
superclass::method
In Java code, you use: superclass.method The equivalent MATLAB operation is method@superclass.
Other Differences
In MATLAB classes, there is no equivalent to C++ templates or Java generics. However, MATLAB is weakly typed and it is possible to write functions and classes that work with different types of data. MATLAB classes do not support overloading functions using different signatures for the same function name.
Modifying Objects
MATLAB classes can define public properties, which you can modify by explicitly assigning values to those properties on a given instance of the class. However, only classes derived from the handle class exhibit reference behavior. Modifying a property value on an instance of a value classes (classes not derived from handle), changes the value only within the context in which the modification is made. The sections that follow describe this behavior in more detail.
3-51
Passing Objects to Functions

MATLAB passes all variables by value. When you pass an object to a function, MATLAB copies the value from the caller into the parameter variable in the called function. However, MATLAB supports two kinds of classes that behave differently when copied: Handle classes a handle class instance variable refers to an object. A copy of a handle class instance variable refers to the same object as the original variable. If a function modifies a handle object passed as an input argument, the modification affects the object referenced by both the original and copied handles. Value classes the property data in an instance of a value class are independent of the property data in copies of that instance (although, a value class property could contain a handle). A function can modify a value object that is passed as an input argument, but this modification does not affect the original object. See Comparing Handle and Value Classes on page 5-2 for more information on the behavior and use of both kinds of classes. Passing Value Objects. When you pass a value object to a function, the function creates a local copy of the argument variable. The function can modify only the copy. If you want to modify the original object, return the modified object and assign it to the original variable name. For example, consider the value class, SimpleClass :
classdef SimpleClass properties Color end methods function obj = SimpleClass(c) if nargin > 0 obj.Color = c; end end end
3-52
end
Create an instance of SimpleClass, assigning a value of red to its Color property:

obj = SimpleClass('red');
Pass the object to the function g, which assigns blue to the Color property:
function y = g(x) x.Color = 'blue'; y = x; end y = g(obj);
The function g modifies its copy of the input object and returns that copy, but does not change the original object.
y.Color ans = blue obj.Color ans = red
If the function g did not return a value, the modification of the object Color property would have occurred only on the copy of obj within the function workspace. This copy would have gone out of scope when the function execution ended. Overwriting the original variable actually replaces it with a new object:
obj = g(obj);
3-53
Passing Handle Objects. When you pass a handle to a function, the function makes a copy of the handle variable, just like when passing a value object. However, because a copy of a handle object refers to the same object as the original handle, the function can modify the object without having to return the modified object. For example, suppose you modify the SimpleClass class definition to make a class derived from the handle class:
classdef SimpleHandleClass < handle properties Color end methods function obj = SimpleHandleClass(c) if nargin > 0 obj.Color = c; end end end end
Create an instance of SimpleHandleClass, assigning a value of red to its Color property:

obj = SimpleHandleClass('red');
Pass the object to the function g, which assigns blue to the Color property:
y = g(obj);
The function g sets the Color property of the object referred to by both the returned handle and the original handle:
y.Color ans = blue obj.Color
3-54
ans = blue
The variables y and obj refer to the same object:

y.Color = 'yellow'; obj.Color ans = yellow
The function g modified the object referred to by the input argument (obj) and returned a handle to that object in y. MATLAB Passes Handles by Value. A handle variable is a reference to an object. MATLAB passes this reference by value. Handles do not behave like references in C++. If you pass an object handle to a function and that function assigns a different object to that handle variable, the variable in the caller is not affected. For example, suppose you define a function g2:
function y = g2(x) x = SimpleHandleClass('green'); y = x; end
Pass a handle object to g2:

obj = SimpleHandleClass('red'); y = g2(obj); y.Color ans = green obj.Color ans =
3-55
red
The function overwrites the handle passed in as an argument, but does not overwrite the object referred to by the handle. The original handle obj still references the original object.
Common Object-Oriented Techniques

This table provides links to sections that discuss object-oriented techniques commonly used by other object-oriented languages. Technique Operator overloading Multiple inheritance Subclassing Destructor Data member scoping Packages (scoping classes) Named constants Enumerations Static methods Static properties How to Use in MATLAB Implementing Operators for Your Class on page 15-31 Subclassing Multiple Classes on page 10-18 Creating Subclasses Syntax and Techniques on page 10-7 Handle Class Destructor on page 5-16 Property Attributes on page 6-8 Packages Create Namespaces on page 4-20 See Properties with Constant Values on page 13-2 and Defining Named Values on page 12-2 Working with Enumerations on page 12-4 Static Methods on page 7-25 Not supported. See persistent variables. For the equivalent of Java static final or C++ static const properties, use Constant properties. See Properties with Constant Values on page 13-2 Class Constructor Methods on page 7-16 No direct equivalent
Constructor Copy constructor
3-56
Technique Reference/reference classes Abstract class/Interface Garbage collection Instance properties Importing classes Events and Listeners
How to Use in MATLAB Comparing Handle and Value Classes on page 5-2 Defining Abstract Classes on page 10-77 Object Lifecycle on page 5-18 Dynamic Properties Adding Properties to an Instance on page 6-27 Importing Classes on page 4-25 Events and Listeners Concepts on page 9-11
3-57
3-58
4
Defining and Organizing Classes
User-Defined Classes on page 4-2 Class Definition on page 4-4 Class Attributes on page 4-6 Expressions in Class Definitions on page 4-9 Organizing Classes in Folders on page 4-15 Class Precedence on page 4-18 Packages Create Namespaces on page 4-20 Importing Classes on page 4-25
In this section... What is a Class Definition on page 4-2 Attributes for Class Members on page 4-2 Kinds of Classes on page 4-3 Constructing Objects on page 4-3 Class Hierarchies on page 4-3
What is a Class Definition

A MATLAB class definition is a template whose purpose is to provide a description of all the elements that are common to all instances of the class. Class members are the properties, methods, and events that define the class. MATLAB classes are defined in code blocks, with sub-blocks delineating the definitions of various class members. See classdef Syntax on page 4-4 for details on the classdef block.
Attributes for Class Members

Attributes modify the behavior of classes and the members defined in the class-definition block. For example, you can specify that methods are static or that properties are abstract, and so on. The following sections describe these attributes: Class Attributes on page 4-6 Method Attributes on page 7-5 Property Attributes on page 6-8 Event Attributes on page 9-16 Class definitions can provide information, such as inheritance relationships or the names of class members without actually constructing the class. See Class Metadata on page 14-2.
4-2
See Specifying Attributes on page 4-7 for more on attribute syntax.
Kinds of Classes
There are two kinds of MATLAB classeshandle and value classes. Handle classes create objects that reference the data contained. Copies refer to the same data. Value classes make copies of the data whenever the object is copied or passed to a function. MATLAB numeric types are value classes. See Comparing Handle and Value Classes on page 5-2 for a more complete discussion.
Constructing Objects
For information on class constructors, see Class Constructor Methods on page 7-16 For information on creating arrays of objects, see Creating Object Arrays on page 8-2
Class Hierarchies
For more information on how to define class hierarchies, see Hierarchies of Classes Concepts on page 10-2.
4-3
Class Definition
In this section... classdef Syntax on page 4-4 Examples of Class Definitions on page 4-4
classdef Syntax
Class definitions are blocks of code that are delineated by the classdef keyword at the beginning and the end keyword at the end. Files can contain only one class definition. The following diagram shows the syntax of a classdef block. Only comments and blank lines can precede the classdef key word.
classdef keyword begins denition block.
classdef (ConstructOnLoad = true) PositiveIntegers < Integers & Positives ... end classdef block
Class attribute
Attribute value (logical true)
Class name
Super classes
end keyword terminates denition block.
Examples of Class Definitions

See the following links for examples of classes defined for different purposes: Class to Represent Structured Data on page 2-22
4-4
Class Definition
Class to Implement Linked Lists on page 2-31 Developing Classes Typical Workflow on page 2-11 A Polynomial Class on page 17-2
4-5
Class Attributes
In this section... Table of Class Attributes on page 4-6 Specifying Attributes on page 4-7
Table of Class Attributes

All classes support the attributes listed in the following table. Attributes enable you to modify the behavior of class. Attribute values apply to the class defined within the classdef block. Attribute Name
Abstract
Class
logical
Description If specified as true, this class is an abstract class (cannot be instantiated). See Defining Abstract Classes on page 10-77 for more information. List classes that can subclass this class. Specify subclasses as meta.class objects in the form: A single meta.class object A cell array of meta.class objects. An empty cell array, {}, is the same as a Sealed class (no subclasses). Specify meta.class objects using the ?ClassName syntax only. See Controlling Allowed Subclasses on page 10-20 for more information.
(default =
false) AllowedSubclasses meta.class
object or cell array of

meta.class
objects
ConstructOnLoad
logical
(default =
false)
If true, MATLAB calls the class constructor when loading an object from a MAT-file. Therefore, you must implement the constructor so it can be called with no arguments without producing an error. See Calling Constructor When Loading on page 11-25
4-6
Class Attributes
(Continued) Attribute Name

HandleCompatible
Class
logical
Description If specified as true, this class can be used as a superclass for handle classes. All handle classes are HandleCompatible by definition. See Supporting Both Handle and Value Subclasses on page 10-34 for more information. If true, this class does not appear in the output of the superclasses or help functions. Use this attribute to establish a precedence relationship among classes. Specify a cell array of meta.class objects using the ? operator. The built-in classes double, single, char, logical, int64, uint64, int32, uint32, int16, uint16, int8, uint8, cell, struct, and function_handle are always inferior to user-defined classes and do not show up in this list. See Class Precedence on page 4-18 If true, this class can not be subclassed.
(default = false)for value classes

Hidden logical
(default =
false) InferiorClasses cell
(default = {})
Sealed
logical
(default =
false)
Attributes are specified for class members in the classdef, properties, methods, and events definition blocks. The particular attribute setting applies to all members defined within that particular block. This means that, for example, you might use multiple properties definition blocks so you can apply different attribute setting to different properties.
Superclass Attributes Are Not Inherited

Class attributes are not inherited, so superclass attributes do not affect subclasses.
4-7
Attribute Syntax
Specify class attribute values in parentheses, separating each attribute name/attribute value pair with a comma. The attribute list always follows the classdef or class member key word, as shown below:
classdef (attribute-name = expression, ...) ClassName properties (attribute-name = expression, ...) ... end methods (attribute-name = expression, ...) ... end events (attribute-name = expression, ...) ... end end
See Expressions in Attribute Specifications on page 4-10 for more information.
4-8
Expressions in Class Definitions

In this section... Basic Knowledge on page 4-9 Where to Use Expressions in Class Definitions on page 4-9 How MATLAB Evaluates Expressions on page 4-11 When MATLAB Evaluates Expressions on page 4-11 Samples of Expression Evaluation on page 4-12
Basic Knowledge
The material presented in this section builds on an understanding of the following information: Operators and Elementary Operations Properties on page 3-10 Specifying Attributes on page 3-23
Where to Use Expressions in Class Definitions

An expression used in a class definition can be any valid MATLAB statement that evaluates to a single array. Use expressions to define property default values and in attribute specifications. Here are some examples used in a class definition:
classdef MyClass (Sealed = true) % Logical value sets attribute properties (Constant = true) CnstProp = 2^.5; end properties Prop1 = MyClass.setupAccount; Prop2 = MyConstants.Minimum; Prop4 = AccountManager; end end % Static method of this class % Constant property from another class % A class constructor
Prop3 = MyConstants.Rate*MyClass.CnstProp % Constant property from this class
4-9
MATLAB does not call property set methods when assigning the result of default value expressions to properties. (See Property Access Methods on page 6-14 for information about these special methods.)
Expressions in Attribute Specifications

Class definitions specify attribute values using an expression that assigns the desired value to the named attribute. For example, this assignment makes MyClass sealed (cannot be subclassed).
classdef MyClass (Sealed = true)
It is possible to use a MATLAB expression on the right side of the equals sign (=) as long as it evaluates to logical true or false. However, this expression cannot use any definitions in its own file, including any constant properties, static methods, and local functions. While it is possible to use conditional expressions to specify attribute values, doing so can cause the class definition to change based on external conditions. See Specifying Attributes on page 3-23 for more information on attribute syntax.
Expressions in Default Property Specifications

Property definitions allow you to specify default values for properties using any expression that has no reference to variables. For example, Myclass defines a constant property (Deg2Rad) and uses it in an expression that defines the default value of another property (PropA). The default value expression also uses a static method (getAngle) defined by the class:
classdef MyClass properties (Constant = true) Deg2Rad = pi/180; end properties PropA = sin(Deg2Rad*MyClass.getAngle[1 0],[0 1]); end ... methods (Static = true) r = getAngle(vx,vy)
4-10
... end end end
Expressions in Class Methods

Expression in class methods execute like expressions in any function MATLAB evaluates an expression within the functions workspace only when the method executes. Therefore, expressions used in class methods are not considered part of the class definition and are not discussed in this section.
How MATLAB Evaluates Expressions

MATLAB evaluates the expressions used in the class definition without any workspace. Therefore, these expressions cannot reference variables of any kind. MATLAB evaluates expressions in the context of the class file, so these expressions can access any functions, static methods, and constant properties of other classes that are on your path at the time MATLAB initializes the class. Expressions defining property default values can access constant properties defined in their own class.
When MATLAB Evaluates Expressions

MATLAB evaluates the expressions in class definitions only when the class is initialized. Initialization occurs before the class is first used. After initialization, the values returned by these expressions are part of the class definition and are constant for all instances of the class. Each instance of the class uses the results of the initial evaluation of the expressions without reevaluation. If you clear a class, then MATLAB reinitializes the class by reevaluating the expressions that are part of the class definition. (see Modifying and Reloading Classes on page 3-43)
4-11
Samples of Expression Evaluation

The following example shows how value and handle object behave when assigned to properties as default values. Suppose you have the following classes. ContClass defines the object that is created as a default property value, and ClassExp has a property that contains a ContClass object:
classdef ContClass properties TimeProp = datestr(now); % Assign current date and time end end classdef ClassExp properties ObjProp = ContClass; end end
MATLAB creates an instance of the ContClass class when the ClassExp class is first used. MATLAB initializes both classes at this time. All instances of ClassExp include a copy of this same instance of ContClass.
a = ClassExp; a.ObjProp.TimeProp ans = 08-Oct-2003 17:16:08
The TimeProp property of the ContClass object contains the date and time when MATLAB initialized the class. Creating additional instances of the ClassExp class shows that the date string has not changed:
b = ClassExp; b.ObjProp.TimeProp ans = 08-Oct-2003 17:16:08
4-12
Because this example uses a value class for the contained object, each instance of the ClassExp has its own copy of the object. For example, suppose you change the value of the TimeProp property on the object contained by ClassExp objectb:
b.ObjProp.TimeProp = datestr(now) ans = 08-Oct-2003 17:22:49
The copy of the object contained by object a is unchanged:

a.ObjProp.TimeProp ans = 08-Oct-2003 17:16:08
Now consider the difference in behavior if the contained object is a handle object:
classdef ContClass < handle properties TimeProp = datestr(now); end end
Creating two instances of the ClassExp class shows that MATLAB created an object when it initialized the ContClass and used a copy of the object handle for each instance of the ClassExp class. This means there is one ContClass object and the ObjProp property of each ClassExp object contains a copy of its handle. Create an instance of the ClassExp class and note the time of creation:
a = ClassExp; a.ObjProp.TimeProp ans =
4-13
08-Oct-2003 17:46:01
Create a second instance of the ClassExp class. The ObjProp contains the handle of the same object:
b = ClassExp; b.ObjProp.TimeProp ans = 08-Oct-2003 17:46:01
Reassign the value of the contained objects TimeProp property:

b.ObjProp.TimeProp = datestr(now); b.ObjProp.TimeProp ans = 08-Oct-2003 17:47:34
Because the ObjProp property of object b contains a handle to the same object as the ObjProp property of object a, the value of the TimeProp property has changed on this object as well:
a.ObjProp.TimeProp ans = 08-Oct-2003 17:47:34
See Comparing Handle and Value Classes on page 5-2 for more information on handle and value classes.
4-14
Organizing Classes in Folders

In this section... Options for Class Folders on page 4-15 @-Folders on page 4-15 Path Folders on page 4-16 Access to Functions Defined in Private Folders on page 4-16 Class Precedence and MATLAB Path on page 4-16
Options for Class Folders

There are two types of folders that can contain class definitions. Each behave differently in a number of respects. @-folders Folder name begins with @ and is not on the MATLAB path, but its parent folder is on the path. Use this type of folder when you want to use multiple files for one class definition. path folders Folder name does not use an @ character and is itself on the MATLAB path. Use this type of folder when you want multiple classes in one folder.
@-Folders
An @-folder must be contained by a path folder, but is not itself on the MATLAB path. Place the class definition file inside the @-folder, which can also contain separate method files. The class definition file must have the same name as the @-folder (without the @-sign) and the class definition (beginning with the classdef key word) must appear in the file before any other code (white space and comments do not constitute code). Define only one class per folder. All files must have a .m extension (for example, @MyClass/MyClass.m, @MyClass/myMethod.m, and so on). You must use an @-folder if you want to use more than one file for your class definition. Methods defined in separate files match the file name to the function name and must be declared in the classdef file. See Methods In Separate Files on page 3-16 for more information.
4-15
Path Folders
You can locate class definition files in folders that are on the MATLAB path. These classes are visible on the path like any ordinary function. Class definitions placed in path folders behave like any ordinary function with respect to precedencethe first occurrence of a name on the MATLAB path takes precedence over all subsequent occurrences. The name of the file must match the name of the class, as specified with the classdef key word. Using a path folder eliminates the need to create a separate @-folder for each class. However, the entire class definition, including all methods, must be contained within a single file (for example, MyClass1.m, MyClass2.m, and so on). See the path function for information about the MATLAB path.
Access to Functions Defined in Private Folders

Private folders contain functions that are accessible only from functions defined in folders immediately above the private folder (See Private Functions for more information). If a class folder contains a private folder, only the class (or classes) defined in that folder can access functions defined in the private folder. Subclasses do not have access to superclass private functions. If you want a subclass to have access to the private functions of the superclass, define the private functions as protected methods of the superclass (that is, in a methods block with the Access attribute defined a protected).
No Class Definitions in Private Folders

You cannot put class definitions in private folders because doing so would not meet the requirements for @ or path folders.
Class Precedence and MATLAB Path

When multiple class definition files with the same name exist, the precedence of a given file is determined by its location on the MATLAB path. All class definition files before it on the path (whether in an @-folder or not) take precedence and it takes precedence over all class definition files occurring later on the path.
4-16
For example, consider a path with the following folders, containing the files indicated:
fldr1/foo.m fldr2/foo.m fldr3/@foo/foo.m fldr4/@foo/bar.m fldr5/foo.m % % % % % defines defines defines defines defines class foo function foo class foo method bar class foo
The MATLAB language applies the logic in the following list to determine which version of foo to call: Class fldr1/foo.m takes precedence over the class fldr3/@foo because it is before fldr3/@foo on the path. Class fldr3/@foo takes precedence over function fldr2/foo.m because it is a class in an @-folder and fldr2/foo.m is not a class (@-folder classes take precedence over functions). Function fldr2/foo.m takes precedence over class fldr5/foo.m because it comes before class fldr5/foo.m on the path and because class fldr5/foo.m is not in an @-folder. Classes not defined in @-folder abide by path order with respect to functions. Class fldr3/@foo takes precedence over class fldr4/@foo; therefore, the method bar is not recognized as part of the foo class (which is defined only by fldr3/@foo). If fldr3/@foo/foo.m does not contain a classdef keyword (i.e., it is a MATLAB class prior to Version 7.6), then fldr4/@foo/bar.m becomes a method of the foo class defined in fldr3/@foo.
Previous Behavior of Classes Defined in @-Folders

In MATLAB Versions 5 through 7, @-folders do not shadow other @-folders having the same name, but residing in later path folders. Instead, the class is defined by the combination of methods from all @-folders having the same name. This is no longer true. Note that for backward compatibility, classes defined in @-folders always take precedence over functions and scripts having the same name, even those that come before them on the path.
4-17
Class Precedence
InferiorClasses Attribute
You can specify the relative precedence of user-defined classes using the class InferiorClasses attribute. Assign a cell array of class names (represented as meta.class objects) to this attribute to specify classes that are inferior to the class you are defining. For example, the following classdef declares that myClass is dominant over class1 and class2.
classdef (InferiorClasses = {?class1,?class2}) myClass ... end
The ? operator combined with a class name creates a meta.class object. This syntax enables you to create a meta.class object without requiring you to construct an actual instance of the class. MATLAB built-in classes are always inferior to user-defined classes and should not be used in this list. The built-in classes include: double, single, char, logical, int64, uint64, int32, uint32, int16, uint16, int8, uint8, cell, struct, and function_handle.
Dominant Class
MATLAB uses class dominance when evaluating expressions involving objects of more than one class. The dominant class determines: The methods of which class MATLAB calls when more than one class defines methods with the same names. The class of arrays that are formed by combining objects of different classes, assuming MATLAB can convert the inferior objects to the dominant class. See Concatenating Objects of Different Classes on page 8-13 for more information on creating object arrays.
4-18
Class Precedence
More Information
See Determining Which Method Is Invoked on page 7-9 for more on how the MATLAB classes dispatch when evaluating expressions containing objects. See Class Precedence and MATLAB Path on page 4-16 for information on how the location of a class definition on the MATLAB path determines its precedence. See Class Metadata on page 14-2 for information on meta-class objects.
No Attribute Inheritance
Subclasses do not inherit a superclass InferiorClasses attribute. Only instances of the classes specified in the subclass InferiorClasses attribute are inferior to subclass objects.
4-19
Packages Create Namespaces

In this section... Internal Packages on page 4-20 Package Folders on page 4-20 Referencing Package Members Within Packages on page 4-21 Referencing Package Members from Outside the Package on page 4-22 Packages and the MATLAB Path on page 4-23
Internal Packages
MathWorks reserves the use of packages named internal for utility functions used by internal MATLAB code. Functions that belong to an internal package are intended for MathWorks use only. Using functions or classes that belong to an internal package is strongly discouraged. These functions and classes are not guaranteed to work in a consistent manner from one release to the next. Any of these functions and classes might be removed from the MATLAB software in any subsequent release without notice and without documentation in the product release notes.
Package Folders
Packages are special folders that can contain class folders, function and class definition files, and other packages. Packages define the scope of the contents of the package folder (that is, a namespace within which names must be unique). This means function and class names need to be unique only within the package. Using a package provides a means to organize classes and functions and to select names for these components that other packages can reuse. Note Packages are not supported for classes created prior to MATLAB Version 7.6 (i.e., classes that do not use classdef). Package folders always begin with the + character. For example,
4-20
+mypack +mypack/pkfcn.m % a package function +mypack/@myClass % class folder in a package
The top-level package folders parent folder must be on the MATLAB path.
Listing the Contents of a Package

List the contents of a package using the help command:
help event Contents of event: EventData PropertyEvent listener proplistener -
event.EVENTDATA Base class for event da event.PROPERTYEVENT Event data for obje event.LISTENER Listener object event.PROPLISTENER Listener object for
You can also use the what command:

what event Classes in directory Y:xxx\matlab\toolbox\matlab\lang\+event EventData PropertyEvent listener proplistener
Referencing Package Members Within Packages

All references to packages, functions, and classes in the package must use the package name prefix, unless you import the package. (See Importing Classes on page 4-25.) For example, call a package function with this syntax:
z = mypack.pkfcn(x,y);
Note that definitions do not use the package prefix. For example, the function definition line of the pkfcn.m function would include only the function name:
function z = pkfcn(x,y)
Similarly, a package class would be defined with only the class name:
classdef myClass
4-21
but would be called with the package prefix:

obj = mypack.myClass(arg1,arg2,...);
Calling class methods does not require the package name because you have an instance of the class:
obj.myMethod(arg) or myMethod(obj,arg)
A static method requires the full class name:

mypack.myClass.stMethod(arg)
Referencing Package Members from Outside the Package

Because functions, classes, and other packages contained in a package are scoped to that package, to reference any of the package members, you must prefix the package name to the member name, separated by a dot. For example, the following statement creates an instance of myClass, which is contained in mypack package.
obj = mypack.myClass;
Accessing Class Members Various Scenarios

This section shows you how to access various package members from outside a package. Suppose you have a package mypack with the following contents:
+mypack +mypack/myfcn.m +mypack/@myfirstclass +mypack/@myfirstclass/myfcn.m +mypack/@myfirstclass/otherfcn.m +mypack/@myfirstclass/myfirstclass.m +mypack/@mysecondclass +mypack/@mysecondclass/mysecondclass.m +mypack/+mysubpack +mypack/+mysubpack/myfcn.m
Invoke the myfcn function in mypack:
4-22
mypack.myfcn(arg)
Create an instance of each class in mypack:

obj1 = mypack.myfirstclass; obj2 = mypack.mysecondclass(arg);
Invoke the myfcn function in mysubpack:

mypack.mysubpack.myfcn(arg1,arg2);
If mypack.myfirstclass has a method called myfcn, it is called as any method call on an object:
obj = mypack.myfirstclass; myfcn(obj,arg);
If mypack.myfirstclass has a property called MyProp, it can be assigned using dot notation and the object:
obj = mypack.myfirstclass; obj.MyProp = some_value;
Packages and the MATLAB Path

You cannot add package folders to the MATLAB path, but you must add the packages parent folder to the path. Even if a package folder is the current folder, its parent folder must still be on the MATLAB path or the package members are not accessible. Package members remain scoped to the package even if the package folder is the current folder. You must, therefore, always refer to the package members using the package name. Package folders do not shadow other package folders that are positioned later on the path, unlike classes, which do shadow other classes.
Resolving Redundant Names

Suppose a package and a class have the same name. For example:
fldr1/+foo
4-23
fldr2/@foo/foo.m
A call to which foo returns the path to the executable class constructor:
>> which foo fldr2/@foo/foo.m
A function and a package can have the same name. However, a package name by itself is not an identifier so if a redundant name occurs alone, it identifies the function. Executing a package name alone returns an error.
Package Functions vs. Static Methods

In cases where a package and a class have the same name, a static method takes precedence over a package function. For example:
fldr1/+foo/bar.m % bar is a function in package foo fldr2/@foo/bar.m % bar is a static method of class foo
A call to which foo.bar returns the path to the static method:

>> which foo.bar fldr2/@foo/bar.m
In cases where a path folder contains both package and class folders with the same name, the class static method takes precedence over the package method:
fldr1/@foo/bar.m % bar is a static method of class foo fldr1/+foo/bar.m % bar is a function in package foo
A call to which foo.bar returns the path to the static method:

>> which foo.bar fldr1/@foo/bar.m
4-24
Importing Classes
Importing Classes
In this section... Related Information on page 4-25 Syntax for Importing Classes on page 4-25
Related Information
See Packages Create Namespaces on page 4-20 for information about packages.
Syntax for Importing Classes

You can import classes into a function to simplify access to class members. For example, suppose there is a package that contains a number of classes, but you need to use only one of these classes in your function, or perhaps even just a static method from that class. You can use the import command as follows:
function myFunc import pkg.cls1 obj = cls1(arg,...); % call cls1 constructor obj.Prop = cls1.StaticMethod(arg,...); % call cls1 static method end
Note that you do not need to reference the package name (pkg) once you have imported the class (cls1). You can also import all classes in a package using the syntax pkg.*, where * indicates all classes in the package. For example,
function myFunc import pkg.* obj1 = cls1(arg,...); % call pkg.cls1 constructor obj2 = cls2(arg,...); % call pkg.cls2 constructor a = pkgFunction(); % call package function named pkgFunction end
Importing Package Functions

You can use import with package functions:
4-25
function myFunc import pkg.pkfcn pkfcn(arg,...); % call imported package function end
Package Function and Class Method Name Conflict

Suppose you have the following folder organization:
+pkg/timedata.m % package function +pkg/@myclass/myclass.m % class definition file +pkg/@myclass/timedata.m % class method
Now import the package and call timedata on an instance of myclass:

import pkg.* myobj = pkg.myclass; timedata(myobj)
A call to timedata finds the package function, not the class method because MATLAB applies the import and finds pkg.timedata first. Do not use a package in cases where you have name conflicts and plan to import the package.
Clearing Import List

You can not clear the import list from a function workspace. To clear the base workspace only, use:
clear import
4-26
5
Value or Handle Class Which to Use
Comparing Handle and Value Classes on page 5-2 Which Kind of Class to Use on page 5-9 The Handle Superclass on page 5-11 Handle Class Destructor on page 5-16 Finding Handle Objects and Properties on page 5-22 Implementing a Set/Get Interface for Properties on page 5-23 Controlling the Number of Instances on page 5-31
Comparing Handle and Value Classes

In this section... Basic Difference on page 5-2 Why Select Handle or Value on page 5-2 Behavior of MATLAB Built-In Classes on page 5-3 Behavior of User-Defined Classes on page 5-4
Basic Difference
A value class constructor returns an instance that is associated with the variable to which it is assigned. If you reassign this variable, MATLAB creates a copy of the original object. If you pass this variable to a function, the function must return the modified object. A handle class constructor returns a handle object that is a reference to the object created. You can assign the handle object to multiple variables or pass it to functions without causing MATLAB to make a copy of the original object. A function that modifies a handle object passed as an input argument does not need to return the object. Note All handle classes must subclass the abstract handle class. Modifying Objects on page 3-51 compares handle and value object behavior when used as arguments to functions.
Why Select Handle or Value

MATLAB support two kinds of classes handle classes and value classes. The kind of class you use depends on the desired behavior of the class instances and what features you want to use. Use a handle class when you want to create a reference to the data contained in an object of the class, and do not want copies of the object to make copies of the object data. For example, use a handle class to implement an object that
5-2
contains information for a phone book entry. Multiple application programs can access a particular phone book entry, but there can be only one set of underlying data. The reference behavior of handles enables these classes to support features like events, listeners, and dynamic properties. Use value classes to represent entities that do not need to be unique, like numeric values. For example, use a value class to implement a polynomial data type. You can copy a polynomial object and then modify its coefficients to make a different polynomial without affecting the original polynomial. Which Kind of Class to Use on page 5-9 describes how to select the kind of class to use for your application.
Behavior of MATLAB Built-In Classes

If you create an object of the class int32 and make a copy of this object, the result is two independent objects having no data shared between them. The following code example creates an object of class int32 and assigns it to variable a, and then copies it to b. When you raise a to the fourth power and assign the value again to the variable a, MATLAB creates an object with the new data and assigns it to the variable a, overwriting the previous assignment. The value of b does not change.
a = int32(7); b = a; a = a^4; b 7
MATLAB copies the value of a to b, which results in two independent versions of the original object. This behavior is typical of MATLAB numeric classes. Handle Graphics classes return a handle to the object created. A handle is a variable that references an instance of a class. If you copy the handle, you have another variable that refers to the same object. There is still only one version of the object data. For example, if you create a Handle Graphics line object and copy its handle to another variable, you can set the properties of the same line using either copy of the handle.
5-3
x = 1:10; y = sin(x); h1 = line(x,y); h2 = h1; set(h2,'Color','red') % line is red set(h1,'Color','green') % line is green delete(h2) set(h1,'Color','blue')
MATLAB returns an
Invalid or deleted object.
error in this case. If you delete one handle, all copies are now invalid because you have deleted the single object to which all copies point.
Behavior of User-Defined Classes

Value class instances behave like built-in numeric classes and handle class instances behave like Handle Graphics objects, as illustrated in Behavior of MATLAB Built-In Classes on page 5-3.
Value Classes
MATLAB associates objects of value classes with the variables to which you assign them. When you copy a value object, MATLAB also copies the data contained by the object. The new object is independent of changes to the original object. Instances behave like standard MATLAB numeric and struct classes. Each property behaves essentially like a MATLAB array See Memory Allocation for Arrays for more information.
Value Class Behavior

Use value classes when assigning an object to a variable and passing an object to a function must make a copy of the function. Value objects are always associated with one workspace or temporary variable and go out of scope when that variable goes out of scope or is cleared. There are no references to value objects, only copies which are themselves objects.
5-4
For example, suppose you define a polynomial class whose Coefficients property stores the coefficients of the polynomial. Note how copies of these value-class objects are independent of each other:
p = polynomial([1 0 -2 -5]); p2 = p; p.Coefficients = [2 3 -1 -2 -3]; p2.Coefficients ans = 1 0 -2 -5
Creating a Value Class

All classes that are not subclasses of the handle class are value classes. Therefore, the following classdef creates a value class named myValueClass:
classdef myValueClass
...
end
Handle Classes
Objects of handle classes use a handle to reference objects of the class. A handle is a variable that identifies an instance of a class. When you copy a handle object, MATLAB copies the handle, but not the data stored in the object properties. The copy refers to the same data as the original handle. If you change a property value on the original object, the copied object reflects the same change. All handle classes are subclasses of the abstract handle class. In addition to providing handle copy semantics, deriving from the handle class enables your class to: Inherit a number of useful methods (Handle Class Methods on page 5-12) Define events and listeners (Events and Listeners Syntax and Techniques on page 9-18) Define dynamic properties (Dynamic Properties Adding Properties to an Instance on page 6-27)
5-5
Implement Handle Graphics type set and get methods (Implementing a Set/Get Interface for Properties on page 5-23)
Creating a Handle Class

Subclass the handle class explicitly to create a handle class:
classdef myClass < handle
...
end
See The Handle Superclass on page 5-11 for more information on the handle class and its methods.
Subclasses of Handle Classes

If you subclass a class that is itself a subclass of the handle class, your subclass is also a handle class. You do not need to specify the handle superclass explicitly in your class definition. For example, the employee class is a subclass of the handle class:
classdef employee < handle ... end
Create a subclass of the employee class for engineer employees, which is also a handle class. You do not need to specify handle as a superclass in the classdef:
classdef engineer < employee ... end
Handle Class Behavior

A handle is an object that references its data indirectly. When constructing a handle, the MATLAB runtime creates an object with storage for property values and the constructor function returns a handle to this object. When you assign the handle to a variable or when you pass the handle to a function, MATLAB copies the handle, but not the underlying data.
5-6
For example, suppose you have defined a handle class that stores data about company employees, such as the department in which they work:
classdef employee < handle properties Name = '' Department = ''; end methods function e = employee(name,dept) e.Name = name; e.Department = dept; end % employee function transfer(obj,newDepartment) obj.Department = newDepartment; end % transfer end end
The transfer method in the previous code changes the employees department (the Department property of an employee object). In the following statements, e2 is a copy of the handle object e. Notice that when you change the Department property of object e, the property value also changes in object e2.
e = employee('Fred Smith','QE'); e2 = e; % Copy handle object transfer(e,'Engineering') e2.Department ans = Engineering
The variable e2 is an alias for e and refers to the same property data storage as e.
Initializing Properties to Handle Objects

See Initializing Property Values on page 3-10 for information on the differences between initializing properties to default values in the properties block and initializing properties from within the constructor. Also, see Initializing Arrays of Handle Objects on page 8-7 for related information on working with handle classes.
5-7
employee as a Value Class

If the employee class was a value class, then the transfer method would modify only its local copy of the employee object. In value classes, methods like transfer that modify the object must return a modified object to copy over the existing object variable:
function obj = transfer(obj,newDepartment) obj.Department = newDepartment; end
When you call transfer, assign the output argument to create the modified object.
e = transfer(e,'Engineering');
In a value class, the transfer method does not affect the variable e2, which is a different employee object. In this example, having two independent copies of objects representing the same employee is not a good design. Hence, implement the employee class as a handle class.
Deleting Handles
You can destroy handle objects before they become unreachable by explicitly calling the delete function. Deleting the handle of a handle class object makes all handles invalid. For example:
delete(e2) e.Department Invalid or deleted object.
Calling the delete function on a handle object invokes the destructor function or functions for that object. See Handle Class Destructor on page 5-16 for more information.
5-8
Which Kind of Class to Use
Which Kind of Class to Use

In this section... Examples of Value and Handle Classes on page 5-9 When to Use Handle Classes on page 5-9 When to Use Value Classes on page 5-10
Examples of Value and Handle Classes

Handle and value classes are useful in different situations. For example, value classes enable you to create new array classes that have the same semantics as MATLAB numeric classes. A Polynomial Class on page 17-2 and Class to Represent Structured Data on page 2-22 provides examples of value classes. Handle classes enable you to create objects that more than one function or object can share. Handle objects allow more complex interactions among objects because they allow objects to reference each other. Class to Implement Linked Lists on page 2-31 and Developing Classes Typical Workflow on page 2-11 provides examples of a handle class.
When to Use Handle Classes

Use a handle class when: No two instances of a class can have the same state, making it impossible to have exact copies. For example:
A copy of a graphics object (such as a line) has a different position in its parents list of children than the object from which it was copied. Therefore, the two objects are not identical. Nodes in lists or trees having specific connectivity to other nodesno two nodes can have the same connectivity.
5-9
The class represents physical and unique objects like serial ports or printers, in which the entity or state cannot exist in a MATLAB variable. However, a handle to such entity can be a variable. The class defines events and notifies listeners when an event occurs (notify is a handle class method). The class creates listeners by calling the handle class addlistener method. The class subclasses the dynamicprops class (a subclass of handle) so that instances can define dynamic properties. The class subclasses the hgsetget class (a subclass of handle) so that it can implement a graphics object style set/get interface to access property values. You want to create a singleton class or a class in which you track the number of instances from within the constructor. MATLAB software never creates a unique handle without calling the class constructor. A copy of a handle object is not unique because both original and copy reference the same data.
When to Use Value Classes

Value class instances behave like normal MATLAB variables. A typical use of value classes is to define data structures. For example, suppose you want to define a class to represent polynomials. This class can define a property to contain a list of coefficients for the polynomial. It can implement methods that enable you to perform various common operations on the polynomial object. For example, implement addition and multiplication without converting the object to another class. A value class is suitable because you can copy a polynomial object and have two objects that are identical representations of the same polynomial. See Subclassing MATLAB Built-In Types on page 10-43 for more information on value classes.
5-10
The Handle Superclass

In this section... Building on the Handle Class on page 5-11 Handle Class Methods on page 5-12 Relational Methods on page 5-12 Testing Handle Validity on page 5-13 When MATLAB Destroys Objects on page 5-15
Building on the Handle Class

The handle class is an abstract class, which means you cannot create an instance of this class directly. Instead, you use this class as a superclass when you implement your own class. The handle class is the foundation of all classes that are themselves handle classes. When you define a class that is a subclass of handle, you have created a handle class. Therefore, all classes that follow handle semantics are subclasses of the handle class.
Handle Subclasses
There are two subclasses of the handle class that provide additional features when you derive your class from these subclasses: hgsetget Provides set and get methods to access property values. See Implementing a Set/Get Interface for Properties on page 5-23 for information on subclassing hgsetget. dynamicprops Provides the ability to define instance properties. See Dynamic Properties Adding Properties to an Instance on page 6-27 for information on subclassing dynamicprops. Deriving from subclasses of the handle class means that your class is a handle class. It inherits all the handle class methods, plus the special features provided by these subclasses.
5-11
Handle Class Methods

While the handle class defines no properties, it does define the methods discussed in this section. Whenever you create a handle class (that is, subclass the handle class), your subclass inherits these methods. You can list the methods of a class by passing the class name to the methods function:
>> methods('handle') Methods for class handle: addlistener delete eq findobj findprop ge gt isvalid le lt ne notify
Static Methods: empty
Events and Listeners Syntax and Techniques on page 9-18 provides information on how to use the notify and addlistener methods, which are related to the use of events. Creating Subclasses Syntax and Techniques on page 10-7 provides general information on defining subclasses.
Relational Methods
function function function function function function TF TF TF TF TF TF = = = = = = eq(H1,H2) ne(H1,H2) lt(H1,H2) le(H1,H2) gt(H1,H2) ge(H1,H2)
The handle class overloads these functions with implementations that allow for equality tests and sorting on handles. For each pair of input arrays, these functions return a logical array of the same size. Each element is an
5-12
element-wise equality or comparison test result. The input arrays must be the same size or one (or both) can be scalar. The method performs scalar expansion as required.
Testing Handle Validity

Use the isvalid handle class method to determine if you have a valid handle object. For example, in this statement:
B = isvalid(H) B is a logical array in which each element is true if, and only if, the corresponding element of H is a valid handle. B is always the same size as H.
Handle Class or Graphics Object Handle

Use the isa function to determine if a handle is of class handle, or is a Java or Handle Graphics handle. For example, consider the button class, which derives from the handle class:
classdef button < handle properties UiHandle end methods function obj = button(pos) if nargin > 0 if length(pos) == 4 obj.UiHandle = uicontrol('Position',pos,'Style','pushbutton'); else error('Improper position') end end end end end
Create a button object by passing a position vector to the button constructor:

h = button([50 20 50 20]);
5-13
Determine the difference between the graphics object handle (stored in the UiHandle property) and the handle class object, h. Use ishandle to test the validity of Handle Graphics object handles:
% h is a handle object >> isa(h,'handle') ans = 1 % The uicontrol object handle is not a handle object >> isa(h.UiHandle,'handle') ans = 0 % The button object is not a graphics object >> ishandle(h) ans = 0 % The uicontrol is a graphics object handle >> ishandle(h.UiHandle) ans = 1
If you close the figure, the ishandle function determines that the Handle Graphics handle is not valid:
>> close >> ishandle(h.UiHandle) ans = 0 h is still of class handle and is still a valid handle object: >> isa(h,'handle') ans = 1
5-14
>> isvalid(h) ans = 1 h is also of class button: >> isa(h,'button') ans = 1
When MATLAB Destroys Objects

MATLAB destroys objects in the workspace of a function when the function: Reassigns an object variable to a new value Does not use an object variable for the remainder of a function Function execution ends When MATLAB destroys an object, it also destroys values stored in the properties of the object and returns any computer memory associated with the object to MATLAB or the operating system. You do not need to free memory in handle classes. However, there can be other operations that you want to perform when destroying an object. For example, closing a file or shutting down an external program that the object constructor started. You can define a delete method in your handle subclass for these purposes. See Handle Class Destructor on page 5-16 for more information.
5-15
Handle Class Destructor

In this section... Basic Knowledge on page 5-16 Syntax of Class Destructor Method on page 5-16 When to Define a Destructor Method on page 5-17 Destructors in Class Hierarchies on page 5-18 Object Lifecycle on page 5-18 Restrict Explicit Object Deletion on page 5-20 Nondestructor Delete Methods on page 5-21
Basic Knowledge
Terms and Concepts Class destructor a method named delete that MATLAB calls implicitly before destroying an object of the a handle class. User-defined code can also call delete explicitly to destroy a handle object. Nondestructor a method named delete that does not meet the syntax requirements of a valid destructor. Consequently, MATLAB does not call this method implicitly when destroying an object. Table of method attributes: Method Attributes on page 7-5
Syntax of Class Destructor Method

When destroying an object, MATLAB implicitly calls the class destructor method, if the class defines one. Create a destructor method by implementing a method named delete. However, MATLAB recognizes a class method named delete as the class destructor only if you define delete as an ordinary method with the appropriate syntax. To be a valid class destructor, the delete method: Must have one scalar input argument that is an object of the class.
5-16
Must not define output arguments Cannot be Sealed, Static, or Abstract In addition, destructors should not: Throw errors Create new handles to the object being destroyed If you define a delete method that can be called with more than one input argument, or that returns any output arguments, then MATLAB does not recognize that method as the class destructor, and does not call it when destroying an object of the class. Declare delete as an ordinary method:
methods function delete(obj) % obj is always scalar ... end end
Calling Delete on an Array

MATLAB calls the destructor method element-wise on an array of objects. Because MATLAB calls the delete method separately for each element in an object array, each delete method is passed only one scalar argument.
When to Define a Destructor Method

Use a class destructor to perform any necessary cleanup operations before MATLAB destroys an object of the class. For example, suppose an object opens a file for writing and you want to close the file in your delete method. This delete function calls fclose on a file identifier that the objects FileID property stores:
function delete(obj) fclose(obj.FileID); end
5-17
The Filewriter Class on page 2-18 is an example of a class that uses this delete method.
Destructors in Class Hierarchies

If you create a hierarchy of classes, each class can define its own class destructor method. When destroying an object, MATLAB calls the destructor of each class in the hierarchy. Therefore, defining a delete method in a handle subclass does not override the handle class delete method; the subclass delete methods augment the superclass delete methods.
Inheriting a Sealed Delete Method

You cannot define a valid destructor that is Sealed. MATLAB returns an error when you attempt to instantiate a class that defines a Sealed destructor. Normally, declaring a method as Sealed prevents subclasses from overriding that method. However, because destructors must be named delete, an inherited method named delete that is Sealed does not prevent subclasses from defining valid destructors. For example, if a superclass defines a method named delete that is not a valid destructor and is Sealed, then subclasses: Can define valid destructors (which are always named delete). Cannot define methods named delete that are not valid destructors.
Destructors in Heterogeneous Hierarchies

Heterogeneous class hierarchies (matlab.mixin.Heterogeneous) require that all methods to which heterogeneous arrays are passed must be sealed. However, the rule does not apply to class destructor methods. Because destructor methods cannot be sealed, you can define a valid destructor in a heterogeneous hierarchy that is not sealed, but does function as a destructor.
Object Lifecycle
MATLAB invokes the destructor delete method when the lifecycle of an object ends. The lifecycle of an object ends when the object is:
5-18
No longer referenced anywhere Explicitly deleted by calling delete on the handle
Inside a Function
The lifecycle of an object referenced by a local variable or input argument exists from the time the variable is assigned until the time it is reassigned, cleared, or no longer referenced within that function or any handle array. A variable goes out of scope when you explicitly clear it or when its function ends. When a variable goes out of scope, if its value belongs to a handle class that defines a delete method, MATLAB calls that method. MATLAB defines no ordering among variables in a function. Do not assume that MATLAB destroys one value before another value when the same function contains multiple values.
Sequence During Handle Object Destruction

MATLAB invokes the delete methods in the following sequence when destroying an object:
1 The delete method for the class of the object 2 The delete method of each superclass class, starting with the immediate
superclasses and working up the hierarchy to the most general superclasses MATLAB invokes the delete methods of superclasses at the same level in the hierarchy in the order specified in the class definition. For example, the following class definition specifies supclass1 before supclass2 so MATLAB calls the delete function of supclass1 before the delete function of supclass2.
classdef myClass < supclass1 & supclass2
Superclass delete methods cannot call methods or access properties belonging to a subclass. After calling each delete method, MATLAB destroys the property values belonging exclusively to the class whose method was called. The destruction of property values that contain other handle objects can cause MATLAB to
5-19
call the delete methods for those objects, if there are no other references to those objects.
Destruction of Objects with Cyclic References

Consider a set of objects that reference other objects of the set such that the references form a cyclic graph. In this case, MATLAB: Destroys the objects if they are referenced only within the cycle Does not destroy the objects as long as there is an external reference to any of the objects from a MATLAB variable outside the cycle MATLAB destroys the objects in the reverse of the order of construction.
Restrict Explicit Object Deletion

You can destroy handle objects by explicitly calling delete on the object:
delete(obj)
A class can prevent explicit destruction of an object by setting its delete method Access attribute to private. MATLAB issues an error if you explicitly call delete on a handle object whose delete method is private. However, a method of the class can call the private delete method. Similarly, if the class delete method Access attribute has a value of protected, only methods of the class and any subclasses can explicitly delete objects of that class. However, when an objects lifecycle ends, MATLAB calls the objects delete method when destroying the object regardless of methods Access attribute setting. See Object Lifecycle on page 5-18 for information on when MATLAB destroys objects and Sequence During Handle Object Destruction on page 5-19 for information on how MATLAB calls object delete methods.
Inherited Private Delete Methods

Class destructor behavior differs from the normal behavior of an overridden method. MATLAB executes each delete method of each superclass of an object upon destruction, even if that delete method is not public.
5-20
When you explicitly call an objects delete method, MATLAB checks the delete method Access attribute in the class defining the object, but not in the superclasses of the object. Therefore, a superclass with a private delete method does not prevent the destruction of subclass objects. Declaring a private delete method makes most sense for Sealed classes because subclasses can define their own delete methods with public Access.
Nondestructor Delete Methods

A class can implement a method named delete that is not a valid class destructor, and therefore is not called implicitly by MATLAB when destroying an object. In this case, delete behaves like a normal method. For example, if the superclass implements a Sealed method named delete that is not a valid destructor, then MATLAB does not allow subclasses to override this method. A delete method defined by a value class cannot be a class destructor. See Basic Difference on page 5-2 for information on the difference between a value and handle class. See Syntax of Class Destructor Method on page 5-16 for information on how to implement a delete method that is a valid destructor.
5-21
Finding Handle Objects and Properties

In this section... Finding Handle Objects on page 5-22 Finding Handle Object Properties on page 5-22
Finding Handle Objects

The findobj method enables you to locate handle objects that meet certain conditions.
function HM = findobj(H,<conditions>)
The findobj method returns an array of handles matching the conditions specified.
Finding Handle Object Properties

The findprop method returns the meta.property object for the specified object and property.
function mp = findprop(h,'PropertyName')
The findprop method returns the meta.property object associated with the PropertyName property defined by the class of h. The property can also be a dynamic property created by the addprop method of the dynamicprops class. You can use the returned meta.property object to obtain information about the property, such as querying the settings of any of its attributes. For example, the following statements determine that the setting of the AccountStatus propertys Dependent attribute is false.
ba = BankAccount(007,50,'open'); mp = findprop(ba,'AccountStatus'); % get meta.property object mp.Dependent ans = 0
Class Metadata on page 14-2 provides more information on meta-classes.
5-22
Implementing a Set/Get Interface for Properties

In this section... The Standard Set/Get Interface on page 5-23 Subclass hgsetget on page 5-23 Get Method Syntax on page 5-23 Set Method Syntax on page 5-24 Class Derived from hgsetget on page 5-25
The Standard Set/Get Interface

The MATLAB Handle Graphics system implements an interface based on set and get methods. These methods enable you to set or query the values of graphics object properties. The hgsetget subclass of the handle class provides implementations of these methods. Derive your class from hgsetget to obtain similar set and get functionality. Note The set and get methods referred to in this section are different from property set access and property get access methods. See Property Access Methods on page 6-14 for information on property access methods.
Subclass hgsetget
Classes inherit set and get methods from hgsetget:
classdef MyClass < hgsetget
Because hgsetget derives from the handle class, MyClass is also a handle class.
Get Method Syntax

Get the value of an object property using the object handle, h, and the property name:
v = get(h,'PropertyName');
5-23
If you specify an array of handles with a single property name, get returns the current property value for each object in H as a cell array of values, (CV):
CV = get(H,'PropertyName');
The CV array is always a column regardless of the shape of H. When prop is a cell array of string property names and H is an array of handles, get returns a cell array of values where each row in the cell corresponds to an object in H and each column in the cell corresponds to a property in prop. get returns the corresponding property values in an m-by-n cell array, where m = length(H) and n = length(prop)
prop = {'PropertyName1','PropertyName2'}; CV = get(H,prop);
If you specify a handle array, but no property names, get returns a struct array in which each structure in the array corresponds to an object in H. Each field in the structure corresponds to a property defined by the class of H. The value of each field is the value of the corresponding property. If you do not assign an output variable, then H must be scalar.
SV = get(H);
See Using Handle Arrays with Get on page 5-28 for an example.
Set Method Syntax

The set method assigns the value of the specified property for the object with handle H. If H is an array of handles, MATLAB assigns the property value to the named property for each object in the array H.
set(H,'PropertyName',PropertyValue)
You can pass a cell array of property names and a cell array of property values to set:
set(H,{'PropertyName1','PropertyName2'},... {Property1Value,Property2Value})
5-24
If length(H) is greater than one, then the property value cell array can have values for each property in each object. For example, if length(H) is 2 (two object handles), then you can use an expression like this:
set(H,{'PropertyName1','PropertyName2'},... {Property11Value,Property12Value;Property21Value,Property22Value})
The preceding statement is equivalent to the follow two statements:

set(H(1),'PropertyName1',Property11Value,'PropertyName2',Property12Value) set(H(2),'PropertyName1',Property21Value,'PropertyName2',Property22Value)
If you specify a scalar handle, but no property names, set returns a struct array with one field for each property in the class of H. Each field contains an empty cell array.
SV = set(h);
See Class Derived from hgsetget on page 5-25 for an example.
Class Derived from hgsetget

This sample class defines a set/get interface and illustrates the behavior of the inherited methods:
classdef LineType < hgsetget % subclass hgsetget properties Style = '-'; Marker = 'o'; end % Public properties properties (SetAccess = protected) Units = 'points'; end % Protected SetAccess methods function obj = LineType(s,m) if nargin > 0 obj.Style = s; obj.Marker = m; end end% LineType function obj = set.Style(obj,val)
5-25
if ~(strcmpi(val,'-') ||... strcmpi(val,'--') ||... strcmpi(val,'..')) error('Invalid line style ') end obj.Style = val; end % set.Style function obj = set.Marker(obj,val) if ~isstrprop(val,'graphic') error('Marker must be a visible character') end obj.Marker = val; end % set.Marker end % methods end % classdef
Create an instance of the class and save its handle:

h = LineType('--','*');
Query the value of any object property using the inherited get method:
get(h,'Marker') ans = *
Set the value of any property using the inherited set method:
set(h,'Marker','Q')
Property Access Methods Are Called

MATLAB calls any property access methods (set.Style or set.Marker in the LineType class) when you use the set and get methods that are inherited from the hgsetget class:
set(h,'Style','-.-') Error using LineType>LineType.set.Style Invalid line style
5-26
Using the set and get methods that are inherited from hgsetget invokes any existing property access methods that would execute when assigning or querying property values using dot notation:
h.Style = '-.-'; Error using LineType>LineType.set.Style Invalid line style
See Property Access Methods on page 6-14 for more information on property access methods.
Listing All Properties

You can create a struct containing object properties and their current values using get with only a handle array as input. For example, the struct SV contains fields whose names correspond to property names. Each field contains the current value of the respective property.
% Create a LineType object and save its handle h = LineType('--','*'); % Query the property values of object h SV = get(h) SV = Style: '--' Marker: '*' Units: 'points'
Create a struct containing the properties that have public SetAccess using set with an object handle:
% Query setable property values S = set(h) S = Style: {} Marker: {}
5-27
The LineType class defines the Units property with SetAccess = protected. Therefore, S = set(h) does not create a field for this property in the sturct S. set cannot return possible values for the properties.
Using Handle Arrays with Get

Suppose you create an array of LineType objects:
H = [LineType('..','z'),LineType('--','q')] H = 1x2 LineType with properties: : Style Marker Units
When H is an array of handles, get returns a (length(H)-by-1) cell array of property values:
CV = get(H,'Style') CV = '..' '--'
When H is an array of handles and you do not specify a property name, get returns a struct array containing fields with name corresponding to property-names. You must assign the output of get to a variable when H is not scalar.
% Assign output of get for nonscalar H SV = get(H) SV = 2x1 struct array with fields: Style Marker Units
5-28
Get the value of the Marker property from the second array element in the SV struct array:
SV(2).Marker ans = q
Handle, Property Name, and Property Value Arrays

You can pass an array of handles, a cell array of property names, and a cell array of property values to set. The property value cell array must have one row of property values for each object in H and each row must have a value for each property in the property name array:.
H = [LineType('..','z'),LineType('--','q')]; set(H,{'Style','Marker'},{'..','o';'--','x'})
The results of this call to set is:

H(1) ans = LineType with properties: Style: '..' Marker: 'o' Units: 'points' H(2) ans = LineType with properties: Style: '--' Marker: 'x' Units: 'points'
5-29
Customizing the Property List

You can customize the way property lists are displayed by redefining the following methods in your subclass: setdisp Called by set when you call set with no output arguments and a single input parameter containing the handle array. getdisp Called by get when you call get with no output arguments and a single input parameter containing the handle array.
5-30
Controlling the Number of Instances
Controlling the Number of Instances

Limiting Instances
You can limit the number of instances of a class that can exist at any one time. For example, a singleton class can have only one instance and provides a way to access this instance. You can create a singleton class using these elements: A persistent variable to contain the instance A sealed class (Sealed attribute set to true) to prevent subclassing A private constructor (Access attribute set to private) A static method to return the handle to the instance, if it exists, or to create the instance when needed.
Implementing a Singleton Class

The following skeletal class definition shows how you can approach the implementation of a class that allows you to create only one instance at a time:
classdef (Sealed) SingleInstance < handle methods (Access = private) function obj = SingleInstance end end methods (Static) function singleObj = getInstance persistent localObj if isempty(localObj) || ~isvalid(localObj) localObj = SingleInstance; end singleObj = localObj; end end end
The getInstance static method returns a handle to the object created, which the class stores in a persistent variable. getInstance creates an instance only the first time called in a session or when the object becomes invalid. For example:
5-31
sobj = SingleInstance.getInstance sobj = SingleInstance with no properties
As long as sobj exists as a valid handle, calling getInstance returns a handle to the same object. If you delete sobj, then calling getInstance creates an object and returns the handle.
delete(sobj) isvalid(sobj) ans = 0 sobj = SingleInstance.getInstance; isvalid(sobj) ans = 1
5-32
6
Properties Storing Class Data
How to Use Properties on page 6-2 Defining Properties on page 6-5 Property Attributes on page 6-8 Mutable and Immutable Properties on page 6-13 Property Access Methods on page 6-14 Properties Containing Objects on page 6-25 Dynamic Properties Adding Properties to an Instance on page 6-27
How to Use Properties

In this section... What Are Properties on page 6-2 Types of Properties on page 6-3
What Are Properties

Properties encapsulate the data that belongs to instances of classes. Data contained in properties can be public, protected, or private. This data can be a fixed set of constant values, or it can be dependent on other values and calculated only when queried. You control these aspects of property behaviors by setting property attributes and by defining property-specific access methods. See Property Attributes on page 6-8 for a summary of property attributes.
Flexibility of Object Properties

In some ways, properties are like fields of a struct object. However, storing data in an object property provides more flexibility. Properties can: Define a constant value that you cannot change outside the class definition. See Properties with Constant Values on page 13-2 Calculate its value based on the current value of other data. See Property Get Methods on page 6-18 Execute a function to determine if an attempt to assign a value meets a certain criteria. See Property Set Methods on page 6-16 Trigger an event notification when any attempt is made to get or set its value. See Property-Set and Query Events on page 9-14 Restrict access by other code to the property value. See the SetAccess and GetAccess attributes Property Attributes on page 6-8 Control whether its value is saved with the object in a MAT-file. See The Default Save and Load Process on page 11-2
6-2
How to Use Properties
Types of Properties
There are two types of properties: Stored properties Use memory and are part of the object Dependent properties No allocated memory and the get access method calculates the value when queried
Features of Stored Properties

Can assign an initial value in the class definition Property value is stored when you save the object to a MAT-file Can use a set access method to control possible values, but you are not required to use such methods.
When to Use Stored Properties

You want to be able to save the property value in a MAT-file The property value is not dependent on other property values
Features of Dependent Properties

Dependent properties save memory because property values that depend on other values are calculated only when needed.
When to Use Dependent Properties

Define properties as dependent when you want to: Compute the value of a property from other values (for example, you can compute area from Width and Height properties). Provide a value in different formats depending on other values. For example, the size of a push button in values determined by the current setting of its Units property. Provide a standard interface where a particular property is or is not used, depending on other values. For example, different computer platforms can have different components on a toolbar).
6-3
Property Access Methods on page 6-14 provides information on defining property access methods.
6-4
Defining Properties
Defining Properties
In this section... Property Definition Block on page 6-5 Accessing Property Values on page 6-6 Inheritance of Properties on page 6-6 Specifying Property Attributes on page 6-7
Property Definition Block

The following illustration shows a typical property specification. The property and end keywords delineate a block of code that defines properties having the same attribute settings.
properties keyword begins definition block. Attribute specification
properties (SetAccess = protected) Coefficients = [0 0 1]; end

properties block Property name Default value
end keyword terminates definition block.
6-5
Assigning a Default Value

The preceding example shows the Coefficients property specified as having a default value of [0 0 1]. You can initialize property values with MATLAB expressions. However, these expressions cannot refer to the class that you are defining in any way, except to call class static methods. MATLAB executes expressions that create initial property values only when initializing the class, which occurs just before first using the class. See Defining Default Values on page 3-11 for more information about how MATLAB evaluates default value expressions.
Accessing Property Values

Property access syntax is like MATLAB structure field syntax. For example, assume there is a polynomial class called polyno that defines a Coefficients property. If you created a polyno object p:
p = polyno([1 0 -2 -3]); % Create an instance p (this code does not execute)
you can access this property as follows:

c = p.Coefficients; % Assign the current property value to c p.Coefficients = [4 0 -2 3 5]; % Assign new property values
When you access a property, MATLAB performs any operations that the property requires. For example, executing a property set or get access method and triggering property access events. See Implementing a Set/Get Interface for Properties on page 5-23 for information on how to define set and get methods for properties.
Inheritance of Properties
When you derive one class from another class, the derived (subclass) class inherits all the properties of the superclass. In general, subclasses define only properties that are unique to that particular class. Superclasses define properties that more than one subclass use.
6-6
Defining Properties
Specifying Property Attributes

Attributes specified with the properties key word apply to all property definitions that follow in that block. If you want to apply attribute settings to certain properties only, reuse the properties keyword and create another property block for those properties. For example, the following code shows the SetAccess attribute set to private for the IndependentVar and Order properties, but not for the Coefficients property:
properties Coefficients = [0 0 1]; end properties (SetAccess = private) IndependentVar Order = 0; end
These properties (and any others placed in this block) have private set access
6-7
Property Attributes
Table of Property Attributes
All properties support the attributes listed in the following table. Attributes enable you to modify the behavior of properties. Attribute values apply to all properties defined within the properties block that specifies the nondefault values. Attribute Name
AbortSet
Class logical default = false
Description If true, and this property belongs to a handle class, then MATLAB does not set the property value if the new value is the same as the current value. This approach prevents the triggering of property PreSet and PostSet events. If true, the property has no implementation, but a concrete subclass must redefine this property without Abstract being set to true. Abstract properties cannot define set or get access methods. See Property Access Methods on page 6-14. Abstract properties cannot define initial values. See Assigning a Default Value on page 6-6. All subclasses must specify the same values as the superclass for the property SetAccess and GetAccess attributes. Abstract=true use with the class attribute Sealed=false (the default).
Abstract
logical default = false
6-8
Property Attributes

Access
Class enumeration, default = public meta.class object cell array of meta.class objects
Description
public unrestricted access protected access from class or
subclasses
private access by class members only
(not subclasses) List of classes that have get and set access to this property. Specify classes as meta.class objects in the form: A single meta.class object A cell array of meta.class objects. An empty cell array, {}, is the same as private access. See Controlling Access to Class Members on page 10-24 Use Access to set both SetAccess and GetAccess to the same value. Query the values of SetAccess and GetAccess directly (not Access).
Constant
Set to true if you want only one value for this property in all instances of the class: Subclasses inherit constant properties, but cannot change them. Constant properties cannot be Dependent. SetAccess is ignored. See Properties with Constant Values on page 13-2 for more information.
6-9

Dependent
Class logical default = false
Description If false, property value is stored in object. If true, property value is not stored in object. The set and get functions cannot access the property by indexing into the object using the property name. MATLAB does not display in the command window the names and values of Dependent properties that do not define a get method (scalar object display only). Using a Dependent Property on page 2-27 Property Get Methods on page 6-18 Avoiding Property Initialization Order Dependency on page 11-23
GetAccess
enumeration default = public
subclasses
(not from subclasses) List classes that have get access to this property. Specify classes as meta.class objects in the form: A single meta.class object A cell array of meta.class objects. An empty cell array, {}, is the same as private access.
6-10
Property Attributes
(Continued) Attribute Name Class Description See Controlling Access to Class Members on page 10-24 MATLAB does not display in the command window the names and values of properties having protected or private GetAccess or properties whose Hidden attribute is true. The struct function defines fields for all properties when converting objects to structs.
GetObservable
If true, and it is a handle class property, then you can create listeners for access to this property. The listeners are called whenever property values are queried. See Property-Set and Query Events on page 9-14 Determines whether the property should be shown in a property list (e.g., Property Inspector, call to set or get, etc.). MATLAB does not display in the command window the names and values of properties whose Hidden attribute is true or properties having protected or private GetAccess.
Hidden
6-11

SetAccess
Class enumeration default = public
Description
subclasses
(not from subclasses)

immutable property can be set only in the constructor.
See Mutable and Immutable Properties on page 6-13 List classes that have set access to this property. Specify classes as meta.class objects in the form: A single meta.class object A cell array of meta.class objects. An empty cell array, {}, is the same as private access. See Controlling Access to Class Members on page 10-24
SetObservable
If true, and it is a handle class property, then you can create listeners for access to this property. The listeners are called whenever property values are modified. See Property-Set and Query Events on page 9-14 If true, property value is not saved when object is saved to a file. See Understanding the Save and Load Process on page 11-2 for more about saving objects.
Transient
6-12
Mutable and Immutable Properties
Mutable and Immutable Properties

Setting Property Values
The property SetAccess attribute enables you to determine under what conditions code can modify object property values. There are four levels of set access that provide varying degrees of access to object property values: SetAccess = public any code with access to an object can set public property values. There are differences between the behavior of handle and value classes with respect to modifying object properties. See Modifying Objects on page 3-51 for information on these differences. SetAccess = protected only code executing from within class methods or methods of subclasses can set property values. You cannot change the value of an object property unless the class or any of its subclasses defines a method to do so. SetAccess = private only the defining class can set property values. You can change the value of an object property only if the class defines a method to perform this action. SetAccess = immutable only the class constructor can set property values. You cannot change the value of an object property.
6-13

In this section... Property Setter and Getter Methods on page 6-14 Property Set Methods on page 6-16 Property Get Methods on page 6-18 Set and Get Methods for Dependent Properties on page 6-19 Set and Get Method Execution and Property Events on page 6-22 Access Methods and Subscripted Reference and Assignment on page 6-23 Performing Additional Steps with Property Access Methods on page 6-23
Property Setter and Getter Methods

Property access methods execute specific code whenever the associated propertys value is referenced or assigned a new value. These methods enable you to perform a variety of operations: Execute code before assigning property values to perform actions such as:
Impose value range restrictions (Restricting Properties to Specific Values on page 2-25) Check for proper types and dimensions Provide error handling
Execute code before returning the current values of properties to perform actions such as: Calculate the value of properties that do not store values (for an example, see Using a Dependent Property on page 2-27) Change the value of other properties Trigger events (for an example, see Defining and Triggering an Event on page 9-4)
Property access methods execute automatically whenever you query or set the corresponding property values.
6-14
Restrictions on Access Methods

You can define property access methods only: For concrete properties (that is, properties that are not abstract) Within the class that defines the property (unless the property is abstract in that class, in which case the concrete subclass must define the access method). MATLAB has no default set or get property access methods. Therefore, if you do not define property access methods, MATLAB software does not invoke any methods before assigning or returning property values. Once defined, only the set and get methods can set and query the actual property values. See Set Method Behavior on page 6-17 for information on cases where MATLAB does not call property set methods. Note Property set and get access methods are not equivalent to user-callable set and get methods used to access property values from an instance of the class. See Implementing a Set/Get Interface for Properties on page 5-23 for information on user-callable set and get methods.
Access Methods Cannot Call Other Functions to Access Property Values

You can set and get property values only from within your property set or get access method. You cannot call another function from the set or get method and attempt to access the property value from that function. For example, an anonymous function that calls another function to do the actual work cannot access the property value. Similarly, an ordinary access function cannot call another function to access the property value.
Defining Access Methods

Access methods have special names that include the propertys name. Therefore, get.PropertyName executes whenever PropertyName is referenced and set.PropertyName executes whenever PropertyName is assigned a new value.
6-15
Define property access methods in a methods block that specifies no attributes. You cannot call these methods, MATLAB calls them when any code accesses the properties. Therefore, property access methods do not appear in the list of class methods returned by the methods command and are not included in the meta.class objects Methods property. However, the meta.property objects SetMethod property contains a function handle to the propertys set method and the GetMethod property contains a function handle to the propertys get method. For example, if the class myClass defines a set function for its Text property, you can obtain a function handle to this method from the meta.class object:
m = ?myClass; m.Properties{1}.SetMethod % Assuming Text is the first property in the cell array ans = @\mydir\@myClass\myClass.m>myClass.set.Text % This is a function handle
The meta.class object (m) contains meta.property objects corresponding to each class property in its Properties property. This example assumes that the Text property corresponds to the first meta.property object in the cell array of meta.property objects. The order of the class properties in the meta.class Properties property is the same as the order in which the class definition defines the properties. Class Metadata on page 14-2 provides more information on using meta-classes.
Function handles discusses the use of function handles.
Property Set Methods

Property set methods have the following syntax, where PropertyName is the name of the property.
methods % No method attributes function obj = set.PropertyName(obj,value) % Value class end
Here obj is the object whose property is being assigned a value and value is the new value that is assigned to the property.
6-16
Value class set functions must return the object with the new value for the property assigned. Value classes replace the object whose property is being assigned with the object returned by the set method. Handle classes do not need to return the modified object.
methods % No method attributes function set.PropertyName(obj,value) % Handle class end
The property set method can perform actions like error checking on the input value before taking whatever action is necessary to store the new property value.
function obj = set.PropertyName(obj,value) if ~(value > 0) error('Property value must be positive') else obj.PropertyName = value; end end
See Restricting Properties to Specific Values on page 2-25 for an example of a property set method.
Set Method Behavior

If a property set method exists, MATLAB calls it whenever a value is assigned to that property. However, MATLAB does NOT call property set methods in the following cases: Assigning a value to a property from within its own property set method, which prevents recursive calling of the set method Specifying default values in class definitions do not invoke the set method Assigning a property to its default value, which is specified in the class definition Copying a value object (that is, not derived from the handle class). Neither the set or get method is called when copying property values from one object to another.
6-17
Assigning a property value that is the same as the current value when the propertys AbortSet attribute is true does not call the propertys set method. See Aborting Set When Value Does Not Change on page 9-31 for more information on this attribute. When assigning a property value, the calling functions copy of the object that has been passed to the set method reflects the changed value. Therefore, an assignment to even a single property is able to affect the whole object. This behavior enables a set method to change other properties in the object as well as its designated property. For example, a graphics window object can have a Units property and a Size property. Changing the Units property can also require a change to the values of the Size property to reflect the new units.
Property Get Methods

MATLAB calls a propertys get method whenever the property value is queried. For example, passing a property value in the following statement causes the method get.XYData to execute, if it exists.
plot(obj.XYData)
Property get methods have the following syntax, where PropertyName is the name of the property. The function must return the property value.
methods % No method attributes function value = get.PropertyName(obj) end
Get Method Behavior

MATLAB does NOT call property get methods in the following cases: Getting a property value from within its own property get method, which prevents recursive calling of the get method Copying a value object (that is, not derived from the handle class). Neither the set or get method is called when copying property values from one object to another.
6-18
Set and Get Methods for Dependent Properties

Dependent properties do not store data because the value of a dependent property depends on the current state of something else, like a concrete property value. Dependent properties must define a get method to determine the value for the property, when queried. Typically, the property get method queries other property values to determine what value to return for the dependent property. For example, the Account class returns a value for the dependent Balance property that depends on the value of the Currency property. The get.Balance method queries the Currency property before calculating a value for the Balance property. MATLAB calls the get.Balance method when the Balance property is queried. You cannot call get.Balance explicitly. Here is a partial listing of the class showing a dependent property and its get method:
classdef Account properties Currency DollarAmount end properties (Dependent) Balance end ... methods function value = get.Balance(obj) c = obj.Currency; switch c case 'E' v = obj.DollarAmount / 1.3; case 'P' v = obj.DollarAmount / 1.5; otherwise v = obj.DollarAmount; end
6-19
format bank value = v; end end end
Get Method for Dependent Property

One application of a property get method is to determine the value of a property only when you need it, and avoid storing the value. To use this approach, set the propertys Dependent attribute to true:
properties (Dependent = true)
PropertyName
end
Now the get method for the PropertyName property determines the value of that property and assigns it to the object from within the method:
function value = get.PropertyName(obj) value = calculateValue; ... end
The get method calls a function or static method calculateValue to calculate the property value and returns value to the code accessing the property. The property get method can take whatever action is necessary within the method to produce the output value. Using a Dependent Property on page 2-27 provide an example of a property get method.
When to Use Set Methods with Dependent Properties

While a dependent property does not store its value, there are situations in which you might want to define a set method for a dependent property. For example, suppose you have a class that changes the name of a property from OldPropName to NewPropName. You want to continue to allow the use of the old name without exposing it to new users. You can make OldPropName
6-20
a dependent property with set and get methods as show in the following example:
properties NewPropName end properties (Dependent, Hidden) OldPropName end methods function obj = set.OldPropName(obj,val) obj.NewPropName = val; end function value = get.OldPropName(obj) value = obj.NewPropName; end end
There is no memory wasted by storing both old and new property values, and code that accesses OldPropName continues to work as expected. It is sometimes useful for a set method of a dependent property to assign values to other properties of the object. Assignments made from property set methods cause the execution of any set methods defined for those properties. See Using a Dependent Property on page 2-27 for an example.
When to Use Private Set Access with Dependent Properties

If you use a dependent property only to return a value, then do not define a set access method for the dependent property. Instead, set the SetAccess attribute of the dependent property to private. For example, consider the following get method for the MaxValue property:
methods function mval = get.MaxValue(obj) mval = max(obj.BigArray(:)); end end
6-21
This example uses the MaxValue property to return a value that it calculates only when queried. For this application, define the MaxValue property as dependent and private:
properties (Dependent, SetAccess = private) MaxValue end
Set and Get Method Execution and Property Events

MATLAB software generates events before and after set and get operations. You can use these events to inform listeners that property values have been referenced or assigned. The timing of event generation is as follows: PreGet Triggered before calling the property get method PostGet Triggered after the property get method has returned its value If a class computes a property value (Dependent = true), then the behaviors of its set events are like the get events: PreSet Triggered before calling the property set method PostSet Triggered after calling the property set method If a property is not computed (Dependent = false, the default), then the assignment statement with the set method generates the events: PreSet Triggered before assigning the new property value within the set method PostSet Triggered after assigning the new property value within the set method Events and Listeners Concepts on page 9-11 provides general information about events and listeners. Creating Property Listeners on page 9-27 provides information about using property events. The PostSet Event Listener on page 9-48 shows an example of a property listener.
6-22
Create a Property Set Listener on page 9-8 is another example that uses property events.
Access Methods and Subscripted Reference and Assignment

Use subscripting as a way to reference or assign property values (that is, a = obj.prop(6) or obj.prop(6) = a) without interfering with property set and get methods. When using subscripted reference, the get method returns the whole property value and MATLAB accesses the value referenced by subscripting that object. For subscripted assignment, MATLAB: Invokes the get method to get the property value Performs the subscripted assignment into the returned property Passes the new property value to the set method MATLAB always passes scalar objects to set and get methods. When reference or assignment occurs on an object array, the set and get methods are called in a loop. See Assigning to Read-Only Properties Containing Objects on page 6-25 for related information.
Performing Additional Steps with Property Access Methods

Property access methods are useful in cases where you want to perform some additional steps before assigning or returning a property value. For example, the Testpoint class uses a property set method to check the range of a value. It then applies scaling if it is within a particular range, and set it to NaN if it is not. The property get methods applies a scale factor before returning its current value:
classdef Testpoint properties (Dependent)
6-23
expectedResult = []; end properties(Constant) scalingFactor = 0.001; end methods function obj = set.expectedResult(obj,erIn) if erIn >= 0 && erIn <= 100 erIn = erIn.*obj.scalingFactor obj.expectedResult = erIn; else obj.expectedResult = NaN; end end function er = get.expectedResult(obj) er = obj.expectedResult/obj.scalingFactor; end end end
6-24
Properties Containing Objects
Properties Containing Objects

Assigning to Read-Only Properties Containing Objects
When a class defines a property with private or protected SetAccess, and that property contains an object which itself has properties, assignment behavior depends on whether the property contains a handle or a value object: Handle object you can set properties on handle objects contained in read-only properties Value object you cannot set properties on value object contained in read-only properties. These example classes illustrate this assignment behavior: ReadOnlyProps class with two read-only properties. The class constructor assigns a handle object of type HanClass to the PropHandle property and a value object of type ValClass to the PropValue property. HanClass handle class with public property ValClass value class with public property
classdef ReadOnlyProps properties(SetAccess = private) PropHandle PropValue end methods function obj = ReadOnlyProps obj.PropHandle = HanClass; obj.PropValue = ValClass; end end end classdef HanClass < handle properties
6-25
Hprop end end classdef ValClass properties Vprop end end
Create an instance of the ReadOnlyProps class:

a = ReadOnlyProps a = ReadOnlyProps with properties: PropHandle: [1x1 HanClass] PropValue: [1x1 ValClass]
Use the private PropHandle property to set the property of the HanClass object it contains:
class(a.PropHandle.Hprop) ans = double a.PropHandle.Hprop = 7;
Attempting to make an assignment to the value class object property is not allowed:
a.PropValue.Vprop = 11; You cannot set the read-only property 'PropValue' of ReadOnlyProps.
6-26
Dynamic Properties Adding Properties to an Instance

In this section... What Are Dynamic Properties on page 6-27 Defining Dynamic Properties on page 6-28 Responding to Dynamic-Property Events on page 6-30 Defining Property Access Methods for Dynamic Properties on page 6-32 Dynamic Properties and ConstructOnLoad on page 6-33
What Are Dynamic Properties

You can attach properties to objects without defining these properties in the class definition. These dynamic properties are sometimes referred to as instance properties. Use dynamic properties to attach temporary data to objects or assign data that you want to associate with a particular instance of a class, but not all objects of that class. It is possible for more than one program to define dynamic properties on the same object so you must take care to avoid name conflicts.
Characteristics of Dynamic Properties

Once defined, dynamic properties behave much like class-defined properties: Set and query the values of dynamic properties using dot notation (see Assigning Data to the Dynamic Property on page 6-29) MATLAB saves and loads dynamic properties when you save and load the objects to which they are attached (see Saving and Loading Dynamic Properties on page 11-20 and Dynamic Properties and ConstructOnLoad on page 6-33) Define attributes for dynamic property (see Setting Dynamic Property Attributes on page 6-28). Add property set and get access methods (see Defining Property Access Methods for Dynamic Properties on page 6-32)
6-27
Listen for dynamic property events (see Responding to Dynamic-Property Events on page 6-30) Access dynamic property values from object arrays, with restricted syntax (see Object Arrays with Dynamic Properties on page 8-10)
Defining Dynamic Properties

Any class that is a subclass of the dynamicprops class (which is itself a subclass of the handle class) can define dynamic properties using the addprop method. The syntax is:
P = addprop(H,'PropertyName')
where:
P is an array of meta.DynamicProperty objects H is an array of handles PropertyName is the name of the dynamic property you are adding to each
object
Naming Dynamic Properties

Use only valid names when naming dynamic properties (see Variable Names). In addition, do not use names that: Are the same as the name of a class method Are the same as the name of a class event Contain a period (.)
Setting Dynamic Property Attributes

Use the meta.DynamicProperty object associated with the dynamic property to set property attributes. For example:
P.Hidden = true;
Remove the dynamic property by deleting its meta.DynamicProperty object:
6-28
delete(P);
The property attributes Constant and Abstract have no meaning for dynamic properties and setting the value of these attributes to true has no effect.
Assigning Data to the Dynamic Property

Suppose, you are using a predefined set of GUI widget classes (buttons, sliders, check boxes, etc.) and you want to store the location on a grid of each instance of the widget class. Assume the widget classes are not designed to store location data for your particular layout scheme and you want to avoid creating a map or hash table to maintain this information separately. Assuming the button class is a subclass of dynamicprops, you could add a dynamic property to store your layout data. Here is a simple class to create a uicontrol button:
classdef button < dynamicprops properties UiHandle end methods function obj = button(pos) if nargin > 0 if length(pos) == 4 obj.UiHandle = uicontrol('Position',pos,... 'Style','pushbutton'); else error('Improper position') end end end end end
Create an instance of the button class, add a dynamic property, and set its value:
b1 = button([20 40 80 20]); % button class uses HG-type position layout b1.addprop('myCoord'); % Add a dynamic property b1.myCoord = [2,3]; % Set the property value
6-29
You can access the dynamic property just like any other property, but only on the instance on which you defined it:
b1.myCoord ans = 2 3
Responding to Dynamic-Property Events

You can attach listeners to dynamicprops objects to monitor the addition of dynamic properties to the object. You can also monitor the removal of dynamic properties, which occurs when you delete the object. The dynamicprops class defines two events and inherits one from handle: ObjectBeingDestroyed Inherited from the handle class. PropertyAdded Triggered when you add a dynamic property to an object derived from the dynamicprops class. PropertyRemoved Triggered when you delete the meta.DynamicProperty object associated with the dynamic property. Suppose you define a button object, as described in the previous section:
b2 = button([20 40 80 20]);
Create a function to attach listeners to the button object, b2, and a listener callback function:
function listenDynoEvent(obj) addlistener(obj,'PropertyAdded',@eventPR); addlistener(obj,'PropertyRemoved',@eventPR); function eventPR(src,evnt) mc = metaclass(src); fprintf(1,'%s %s \n',mc.Name,'object') fprintf(1,'%s %s \n','Event triggered:',evnt.EventName) end end
6-30
Triggering the PropertyAdded Event

Add the listeners to the button object, b2. Then, add a dynamic property, myCoord.
% add listeners listenDynoEvent(b2) % add dynamic property and save meta.DynamicProperty object mp = b2.addprop('myCoord');
The listener callback function, eventPR, executes and displays the object class and event name:
button object Event triggered: PropertyAdded
Delete the dynamic property by deleting the meta.DynamicProperty object:

delete(mp) button object Event triggered: PropertyRemoved
Obtain the meta.DynamicProperty object for a dynamic property using the handle findprop method. Use findprop if you do not have the object returned by addprop:
mp = findprop(b2,'myCoord');
Dynamic Properties and Ordinary Property Events

Dynamic properties support property set and get events so you can define listeners for these properties. Listeners are bound to the particular dynamic property for which you define them. Therefore, if you delete a dynamic property, and then create another one with the same name, the listeners do not respond to events generated by the new property, even though the property has the same name as the property for which the event was defined. Having a listener defined for a deleted dynamic property does not cause an error, but the listener callback is never executed.
6-31
Property-Set and Query Events on page 9-14 provides more information on how to define listeners for these events.
Defining Property Access Methods for Dynamic Properties

Dynamic properties enable you to add properties to class instances without modifying class definitions. You can also define property set access or get access methods without creating new class methods. See Property Setter and Getter Methods on page 6-14 for more on the purpose and techniques of these methods. Note You can set and get the property values only from within your property access methods. You cannot call another function from the set or get method and attempt to access the property value from that function. Here are the steps for creating a property access method: Define a function that implements the desired operations you want to perform before the property set or get occurs. These methods must have the following signatures: mySet(obj,val) or val = myGet(obj) Obtain the dynamic propertys corresponding meta.DynamicProperty object. Assign a function handle pointing to your set or get property function to the meta.DynamicProperty objects GetMethod or SetMethod property. This function does not need to be a method of the class and you cannot use a naming scheme like set.PropertyName. Instead, use any valid function name. Suppose you want to create a property set function for the button class dynamic property myCoord created previously. Write the function as follows:
function set_myCoord(obj,val) if end obj.myCoord = val; % set property value ~(length(val) == 2) % require two values error('myCoords require two values ')
6-32
end
Because button is a handle class, the property set function does not need to return the object as an output argument. Assign the value to the property if the value is valid. Use the handle class method findprop to get the meta.DynamicProperty object:
mb1 = b1.findprop('myCoord'); mb1.SetMethod = @set_myCoord;
The property set function is now called whenever you set this property:
b1.myCoord = [1 2 3] % length must be two Error using button.set_myCoord myCoords require two values
Dynamic Properties and ConstructOnLoad

Setting a classs ConstructOnLoad attribute to true causes MATLAB to call the class constructor when loading the class. Dynamic properties are saved and restored when loading an object. If you are creating dynamic properties from the class constructor, you can cause a conflict if you also set the classs ConstructOnLoad attribute to true. Heres the sequence: A saved object saves the names and values of properties, including dynamic properties When loaded, a new object is created and all properties are restored to the values at the time the object was saved Then, the ConstructOnLoad attribute causes a call to the class constructor, which would create another dynamic property with the same name as the loaded property (see The Default Save and Load Process on page 11-2 for more on the load sequence) MATLAB prevents a conflict by loading the saved dynamic property, and does not execute addprop when calling the constructor. If it is necessary for you to use ConstructOnLoad and you add dynamic properties from the class constructor (and want the constructors call to addprop to be executed at load time) then set the dynamic propertys
6-33
Transient attribute to true. This setting prevents the property from being saved. For example: classdef (ConstructOnLoad) MyClass < dynamicprops function obj = MyClass P = addprop(obj,'DynProp'); P.Transient = true; ... end end
6-34
7
Methods Defining Class Operations
How to Use Methods on page 7-2 Method Attributes on page 7-5 Ordinary Methods on page 7-7 Class Constructor Methods on page 7-16 Static Methods on page 7-25 Overloading Functions for Your Class on page 7-27 Object Precedence in Expressions Using Operators on page 7-29 Class Methods for Graphics Callbacks on page 7-31
How to Use Methods

In this section... Class Methods on page 7-2 Method Naming on page 7-3
Class Methods
Methods are functions that implement the operations performed on objects of a class. Methods, along with other class members support the concept of encapsulationclass instances contain data in properties and class methods operate on that data. This allows the internal workings of classes to be hidden from code outside of the class, and thereby enabling the class implementation to change without affecting code that is external to the class. Methods have access to private members of their class including other methods and properties. This enables you to hide data and create special interfaces that must be used to access the data stored in objects. See Methods That Modify Default Behavior on page 15-2 for a discussion of how to create classes that modify standard MATLAB behavior. See Class Files on page 3-2 for information on the use of @ and path directors and packages to organize your class files. See Methods In Separate Files on page 3-16 for the syntax to use when defining classes in more than one file.
Kinds of Methods
There are specialized kinds of methods that perform certain functions or behave in particular ways: Ordinary methods are functions that act on one or more objects and return some new object or some computed value. These methods are like ordinary MATLAB functions that cannot modify input arguments. Ordinary methods enable classes to implement arithmetic operators and
7-2
How to Use Methods
computational functions. These methods require an object of the class on which to operate. See Ordinary Methods on page 7-7. Constructor methods are specialized methods that create objects of the class. A constructor method must have the same name as the class and typically initializes property values with data obtained from input arguments. The class constructor method must return the object it creates. See Class Constructor Methods on page 7-16 Destructor methods are called automatically when the object is destroyed, for example if you call delete(object) or there are no longer any references to the object. See Handle Class Destructor on page 5-16 Property access methods enable a class to define code to execute whenever a property value is queried or set. See Property Setter and Getter Methods on page 6-14 Static methods are functions that are associated with a class, but do not necessarily operate on class objects. These methods do not require an instance of the class to be referenced during invocation of the method, but typically perform operations in a way specific to the class. See Static Methods on page 7-25 Conversion methods are overloaded constructor methods from other classes that enable your class to convert its own objects to the class of the overloaded constructor. For example, if your class implements a double method, then this method is called instead of the double class constructor to convert your class object to a MATLAB double object. See Converting Objects to Another Class on page 15-7 for more information. Abstract methods serve to define a class that cannot be instantiated itself, but serves as a way to define a common interface used by a number of subclasses. Classes that contain abstract methods are often referred to as interfaces. See Defining Abstract Classes on page 10-77 for more information and examples.
Method Naming
The name of a function that implements a method can contain dots (for example, set.PropertyName) only if the method is one of the following: Property set/get access method (see Property Access Methods on page 6-14)
7-3
Conversion method that converts to a package-qualified class, which requires the use of the package name (see Packages Create Namespaces on page 4-20) You cannot define property access or conversion methods as local functions, nested functions, or separately in their own files. Class constructors and package-scoped functions must use the unqualified name in the function definition; do not include the package name in the function definition statement. See Defining Methods on page 7-7 for more information on how you can define methods. See Rules for Naming to Avoid Conflicts on page 7-28 for related information.
7-4
Method Attributes
Method Attributes
Table of Method Attributes
All methods support the attributes listed in the following table. Attributes enable you to modify the behavior of methods. For example, you can prevent access to a method from outside the class or enable the method to be invoked without a class instance. Attribute values apply to all methods defined within the methods block that specifies the nondefault values.
methods (attribute1=value1,attribute2=value2,...) ... end
Attribute Name
Abstract
Class
logical Default = false
Description If true, the method has no implementation. The method has a syntax line that can include arguments, which subclasses use when implementing the method: Subclasses are not required to define the same number of input and output arguments. However, subclasses generally use the same signature when implementing their version of the method. The method can have comments after the function line. The method does not contain function or end keywords, only the function syntax (e.g., [a,b] = myMethod(x,y))
Access
enumeration, default =
public
Determines what code can call this method: public Unrestricted access protected Access from methods in class or subclasses private Access by class methods only (not from subclasses) List classes that have access to this method. Specify classes as meta.class objects in the form:
meta.class object cell array of

meta.class
objects
7-5
Class
Description
Hidden logical Default = false
A single meta.class object A cell array of meta.class objects. An empty cell array, {}, is the same as private access.
See Controlling Access to Class Members on page 10-24 When false, the method name shows in the list of methods displayed using the methods or methodsview commands. If set to true, the method name is not included in these listings and ismethod does return true for this method name. . If true, the method cannot be redefined in a subclass. Attempting to define a method with the same name in a subclass causes an error. Specify as true to define a method that does not depend on an object of the class and does not require an object argument. Use the class name to call the method: classname.methodname or an instance of the class:
obj.methodname
Sealed
logical Default = false logical Default = false
Static
Static Methods on page 7-25 provides more information.
7-6
Ordinary Methods
Ordinary Methods
In this section... Defining Methods on page 7-7 Determining Which Method Is Invoked on page 7-9 Specifying Precedence on page 7-13 Controlling Access to Methods on page 7-13 Invoking Superclass Methods in Subclass Methods on page 7-14 Invoking Built-In Functions on page 7-15
Defining Methods
You can specify methods: Inside of a class definition block In a separate file in the class @-folder
Methods Inside classdef Block

This example shows the definition of a method (the compute function in this example) within the classdef and methods blocks:
classdef ClassName methods (AttributeName = value,...) function x = compute(obj,inc) x = obj.y + inc; end % compute method ... end % methods block ... end % classedf
7-7
Note Nonstatic methods must include an explicit object variable in the function definition. The MATLAB language does not support an implicit reference in the method function definition. Either of the following statements is correct syntax for calling a method where obj is an object of the class defining the compute method:
obj.compute(inc) compute(obj,inc)
See also Dot Notation vs. Function Notation on page 7-10. Method attributes apply only to that particular methods block, which is terminated by the end statement.
Methods in Separate Files

You can define class methods in separate files within the class @-folder. In this case, create a function in a separate file having the same name as the function (i.e., functionname.m). If you want to specify attribute values for that method, you must declare the method signature within a methods block in the classdef block . For example:
classdef myClass methods (AttributeName = value,...) tdata = testdata(obj,arg1,arg2) ... end % methods ... end % classdef
Do not use methods blocks in the separate files. Define the method as a function. Using the example above, the file testdata.m, must contain the definition of the testdata function. Note that the signatures must match.
function tdata = testdata(myClass_object,argument2,argument3) ... end
7-8
Ordinary Methods
The following limitations apply to methods defined in separate files: If you want to specify attributes for a method defined in a separate file, you must declare this method in a methods block (specifying attribute values) within the classdef block. The syntax declared in the methods block (if used) must match the methods function line. The separate file must be in the class @-folder. The constructor method must be defined within the classdef block and, therefore, cannot be in a separate file. (See Class Constructor Methods on page 7-16 for information on this method.) Set and get property access methods must be defined within the classdef block and, therefore, cannot be in separate files. (See Property Access Methods on page 6-14 for information on these methods.)
Determining Which Method Is Invoked

When the MATLAB runtime invokes an ordinary method that has an argument list, it uses the following criteria to determine which method to call The class of the left-most argument whose class is not specified as inferior to any other arguments class is chosen as the dominant class and its method is invoked. If this class does not define the named method, then a function with that name on the MATLAB path is invoked. If no such function exists, MATLAB issues an error indicating that the dominant class does not define the named method.
Dominant Argument
The dominant argument in a methods argument list determines which version of the method or function that the MATLAB runtime calls. Dominance is determined by the relative precedences of the classes of the arguments. In general, user-defined classes take precedence over built-in MATLAB classes. Therefore, the left most argument determines which method to call. However, user-defined classes can specify the relative dominance of specific classes.
7-9
For example, suppose classA defines classB as inferior and suppose both classes define a method called combine. Calling the method with an object of classB and classA:
combine(B,A)
actually calls the combine method of classA because A is the dominant argument. See Specifying Precedence on page 7-13 for information on how to define class precedence.
Dot Notation vs. Function Notation

MATLAB classes support both function and dot notation syntax for calling methods. For example, if setColor is a method of the class of object X, then calling setColor with function notation would be:
X = setColor(X,'red');
The equivalent method call using dot notation is:

X = X.setColor('red')
However, in certain cases, the results for dot notation can differ with respect to how MATLAB dispatching works: If there is an overloaded subsref, it is invoked whenever using dot-notation. That is, the statement is first tested to see if it is subscripted assignment. If there is no overloaded subsref, then setColor must be a method of X. An ordinary function or a class constructor is never called using this notation. Only the argument X (to the left of the dot) is used for dispatching. No other arguments, even if dominant, are considered. Therefore dot notation can call only methods of X; methods of other argument are never called. A Case Where the Result is Different. Here is an example of a case where dot and function notation can give different results. Suppose you have the following classes:
7-10
Ordinary Methods
classA defines a method called methodA that requires an object of classB as one of its arguments classB defines classA as inferior to classB
classdef classB (InferiorClasses = {?classA}) ... end
The methodA method is defined with two input arguments, one of which is an object of classB:
classdef classA methods function methodA(obj,obj_classB) ... end end classB does not define a method with the same name as methodA. Therefore,
the following syntax causes the MATLAB runtime to search the path for a function with the same name as methodA because the second argument is an object of a dominant class. If a function with that name exists on the path, then MATLAB attempts to call this function instead of the method of classA and most likely returns a syntax error.
obj = classA(...); methodA(obj,obj_classB)
Dot notation is stricter in its behavior. For example, this call to methodA:
obj = classA(...); obj.methodA(obj_classB)
can call only methodA of the class of obj.
Referencing Names with ExpressionsDynamic Reference

You can reference an objects properties or methods using an expression in dot-parentheses syntax:
obj.(expression)
7-11
The expression must evaluate to a string that is the name of a property or a method. For example, the following statements are equivalent:
obj.Property1 obj.('Property1')
In this case, obj is an object of a class that defines a property called Property1. Therefore, you can pass a string variable in the parentheses to reference to property:
propName = 'Property1'; obj.(propName)
You can call a method and pass input arguments to the method using another set of parentheses:
obj.(expression)(arg1,arg2,...)
Using this notation, you can make dynamic references to properties and methods in the same way you can create dynamic references to the fields of structs (see Generate Field Names from Variables for information on MATLAB structures). As an example, suppose an object has methods corresponding to each day of the week and these methods have the same names as the days of the week (Monday, Tuesday, and so on). Also, the methods take as string input arguments, the current day of the month (i.e., the date). Now suppose you write a function in which you want to call the correct method for the current day. You can do this using an expression created with the date and datestr functions:
obj.(datestr(date,'dddd'))(datestr(date,'dd'))
The expression datestr(date,'dddd') returns the current day as a string. For example:
datestr(date,'dddd') ans = Tuesday
7-12
Ordinary Methods
The expression datestr(date,'dd') returns the current date as a string. For example:
datestr(date,'dd') ans = 11
Therefore, the expression using dot-parentheses (called on Tuesday the 11th) is the equivalent of:
obj.Tuesday('11')
Specifying Precedence
Class Precedence on page 4-18 provides information on how you can specify the relative precedence of user-define classes.
Controlling Access to Methods

There might be situations where you want to create methods for internal computation within the class, but do not want to publish these methods as part of the public interface to the class. In these cases, you can use the Access attribute to set the access to one of the following options: public Any code having access to an object of the class can access this method (the default). private Restricts method access to the defining class, excluding subclasses. Subclasses do not inherit private methods. protected Restricts method access to the defining class and subclasses derived from the defining class. Subclasses inherit this method. Local and nested functions inside the method files have the same access as the method. Note that local functions inside a class-definition file have private access to the class defined in the same file.
7-13
Invoking Superclass Methods in Subclass Methods

A subclass can override the implementation of a method defined in a superclass. In some cases, the subclass method might need to execute some additional code instead of completely replacing the superclass method. To do this, MATLAB classes can use a special syntax for invocation of superclass methods from a subclass implementation for the same-named method. The syntax to call a superclass method in a subclass class uses the @ symbol:
MethodName@SuperclassName
For example, the following disp method is defined for a Stock class that is derived from an Asset class. The method first calls the Asset class disp method, passing the Stock object so that the Asset components of the Stock object can be displayed. After the Asset disp method returns, the Stock disp method displays the two Stock properties:
classdef Stock < Asset methods function disp(s) disp@Asset(s) % Call base class disp method first fprintf(1,'Number of shares: %g\nShare price: %3.2f\n',... s.NumShares,s.SharePrice); end % disp end end
See The DocStock disp Method on page 18-10 for more information on this example.
Limitations of Use
The following restrictions apply to calling superclass methods. You can use this notation only within: A method having the same name as the superclass method you are invoking A class that is a subclass of the superclass whose method you are invoking
7-14
Ordinary Methods
Invoking Built-In Functions

The MATLAB builtin function enables you to call the built-in version of a function that has been overloaded by a method. See subsref and subsasgn Within Class Methods Built-In Called on page 15-12 for an example.
7-15
Class Constructor Methods

In this section... Rules for Constructors on page 7-16 Related Information on page 7-17 Examples of Class Constructors on page 7-17 Initializing the Object Within a Constructor on page 7-18 No Input Argument Constructor Requirement on page 7-19 Constructing Subclasses on page 7-20 Errors During Class Construction on page 7-22 Basic Structure of Constructor Methods on page 7-23
Rules for Constructors

A constructor method is a special function that creates an instance of the class. Typically, constructor methods accept input arguments to assign the data stored in properties and always return an initialized object. The constructor has the same name as the class. The only output argument from a constructor is the object constructed. The constructor can return only a single argument. Constructors must always return a valid instance of the class. Never return an empty object from a class constructor. If the class being created is a subclass, MATLAB calls the constructor of each superclass class to initialize the object. Implicit calls to the superclass constructor are made with no arguments. If superclass constructors require arguments, you must call them from the subclass constructor explicitly. If your constructor makes an explicit call to a superclass constructor, this call must occur before any other reference to the constructed object. A class does not need to define a constructor method unless it is a subclass of a superclass whose constructor requires arguments. In this case, you must explicitly call the superclass constructor with the required arguments. See Constructing Subclasses on page 7-20
7-16
If a class does not define a constructor, MATLAB supplies a constructor that takes no arguments and returns a scalar object whose properties are initialized to empty or the values specified as defaults in the property definitions. The constructor supplied by MATLAB also calls all superclass constructors with no arguments. If you create a class constructor, you should implement class constructors so that they can be called with no input arguments, in addition to whatever arguments are normally required See No Input Argument Constructor Requirement on page 7-19 and Basic Structure of Constructor Methods on page 7-23. Constructor functions must return an instance of the constructors class. The constructor should avoid assigning to the constructor output argument because subclasses often call a superclass constructor in the process of creating an instance of the subclass. Calls to superclass constructors cannot be conditional. This means superclass construction calls cannot be placed in loops, conditions, switches, try/catch, or nested functions. See Make No Conditional Calls to Superclass Constructors on page 7-20 for more information. Restrict access to constructors using method attributes, as with any method.
Related Information
See Creating Object Arrays on page 8-2 for information on constructing arrays of objects. See Constructor Calling Sequence on page 12-11 for information specific to constructing enumerations.
Examples of Class Constructors

The following links provide access to examples of class constructors: Implementing the BankAccount Class on page 2-13 The Filewriter Class on page 2-18 Simplifying the Interface with a Constructor on page 2-26 Specializing the dlnode Class on page 2-40
7-17
A Class to Manage uint8 Data on page 10-51 Initializing Superclasses from Subclasses on page 10-7 Constructor Arguments and Object Initialization on page 10-10
Initializing the Object Within a Constructor

Constructor functions must return an initialized object as the only output argument. The output argument is created when the constructor executes, before executing the first line of code. For example, the following constructor function can assign the value of the objects property A as the first statement because the object obj has already been assigned to an instance of myClass.
function obj = myClass(a,b,c) obj.A = a;
...
end
You can call other class methods from the constructor because the object is already initialized. The constructor also creates an object whose properties have their default valueseither empty ([]) or the default value specified in the property definition block. See Property Definition Block on page 6-5 for a description of this syntax and see Defining Default Values on page 3-11 for a discussion of how best to define property values. For example, the following code calls the class method CalculateValue to assign the value of the property Value.
function obj = myClass(a,b,c) obj.Value = obj.CalculateValue(a,b);
...
end
Referencing the Object in a Constructor

When initializing the object, for example, by assigning values to properties, you must use the name of the output argument to refer to the object within
7-18
the constructor. For example, in the following code the output argument is obj and the object is reference as obj:
% obj is the object being constructed function obj = myClass(arg) obj.propert1 = arg*10; obj.method1; ... end
No Input Argument Constructor Requirement

There are cases where the constructor must be able to be called with no input argument: When loading objects into the workspace. If the class ConstructOnLoad attribute is set to true, the load function calls the class constructor with no arguments. When creating or expanding an object array such that not all elements are given specific values, the class constructor is called with no arguments to fill in unspecified elements, (for example, x(10,1) = myclass(a,b,c);). In this case, the constructor is called once with no arguments to populate the empty array elements with copies of this one object. See Creating Empty Arrays on page 8-6 for more information. If there are no input arguments, the constructor creates an object using only default properties values. A good practice is to always add a check for zero arguments to the class constructor to prevent an error if either of the two cases above occur:
function obj = myClass(a,b,c) if nargin > 0 obj.A = a; obj.B = b; obj.C = c;
...
end end
See Basic Structure of Constructor Methods on page 7-23 for ways to handle superclass constructors.
7-19
Constructing Subclasses
Subclass constructor functions must explicitly call superclass constructors if the superclass constructors require input arguments. The subclass constructor must specify these arguments in the call to the superclass constructor using the constructor output argument and the returned object must be assigned to the constructor output argument. Here is the syntax:
classdef MyClass < SuperClass function obj = MyClass(arg) obj = obj@SuperClass(ArgumentList); ... end end
The class constructor must make all calls to superclass constructors before any other references to the object, such as assigning property values or calling ordinary class methods. Also, a subclass constructor can call a superclass constructor only once.
Reference Only Specified Superclasses

The constructor cannot call a superclass constructor with this syntax if the classdef does not specify the class as a superclass.
classdef MyClass < SuperClass
MATLAB calls any uncalled constructors in the left-to-right order in which they are specified in the classdef line. MATLAB passes no arguments to these functions.
Make No Conditional Calls to Superclass Constructors

Calls to superclass constructors must be unconditional and you can have only one call for any given superclass. You must initialize the superclass portion of the object by calling the superclass constructors before you can use the object (for example., to assign property values or call class methods). In cases where you need to call superclass constructors with different arguments depending on some condition, you can conditionally build a cell array of arguments and provide one call to the constructor.
7-20
For example, in the following example the superclass shape constructor is called using some default values when the cube constructor has been called with no arguments:
classdef cube < shape properties SideLength = 0; Color = [0 0 0]; end methods function cube_obj = cube(length,color,upvector,viewangle) if nargin == 0 % Provide default values if called with no arguments super_args{1} = [0 0 1]; super_args{2} = 10; else super_args{1} = upvector; super_args{2} = viewangle; end cube_obj = cube_obj@shape(super_args{:}); if nargin > 0 % Use value if provided cube_obj.SideLength = length; cube_obj.Color = color; end ... end ... end
Zero or More Superclass Arguments

If you are calling the superclass constructor from the subclass constructor and you need to support the case where you call the superclass constructor with no arguments, you must explicitly provide for this syntax. Suppose in the case of the cube class example above, all property values in the shape superclass and the cube subclass have initial values specified in the class definitions that create a default cube. Then you could create an instance of cube without specifying any arguments for the superclass or
7-21
subclass constructors. Here is how you can implement this behavior in the cube constructor:
function obj = cube(length,color,upvector,viewangle) if nargin == 0 % Create empty cell array if no input argsuments super_args = {}; else % Use specified argsuments super_args{1} = upvector; super_args{2} = viewangle; end % Call the superclass constructor with the % empty cell array (no arguments) if nargin == 0 % otherwise cell array is not empty cube_obj = cube_obj@shape(super_args{:}); if nargin > 0 cube_obj.SideLength = length; cube_obj.Color = color; end ... end
More on Subclasses
See Creating Subclasses Syntax and Techniques on page 10-7 for information on creating subclasses.
Errors During Class Construction

If an error occurs during the construction of a handle class, the MATLAB class system calls the class destructor on the object along with the destructors for any objects contained in properties and any initialized base classes. See Handle Class Destructor on page 5-16 for information on how objects are destroyed.
7-22
Basic Structure of Constructor Methods

It is important to consider the state of the object under construction when writing your constructor method. Constructor methods can be structured into three basic sections: Pre-initialization Compute arguments for superclass constructors. Object initialization Call superclass constructors. Post initialization Perform any operations related to the subclass, including referencing and assigning to the object, call class methods, passing the object to functions, and so on. This code illustrates the basic operations performed in each section:
classdef myClass < baseClass1 properties ComputedValue end methods function obj = myClass(a,b,c) %%% Pre Initialization %%% % Any code not using output argument (obj) if nargin == 0 % Provide values for superclass constructor % and initialize other inputs a = someDefaultValue; args{1} = someDefaultValue; args{2} = someDefaultValue; else % When nargin ~= 0, assign to cell array, % which is passed to supclass constructor args{1} = b; args{2} = c; end compvalue = myClass.staticMethod(a); %%% Object Initialization %%% % Call superclass constructor before accessing object % You cannot conditionalize this statement
7-23
obj = obj@baseClass1(args{:}); %%% Post Initialization %%% % Any code, including access to object obj.classMethod(...); obj.ComputedValue = compvalue; ... end ... end ... end
See Creating Object Arrays on page 8-2 for information on creating object arrays in the constructor.
7-24
Static Methods
Static Methods
In this section... Why Define Static Methods on page 7-25 Calling Static Methods on page 7-26
Why Define Static Methods

Static methods are associated with a class, but not with specific instances of that class. These methods do not perform operations on individual objects of a class and, therefore, do not require an instance of the class as an input argument, like ordinary methods. Static methods are useful when you do not want to first create an instance of the class before executing some code. For example, you might want to set up the MATLAB environment or use the static method to calculate data needed to create class instances. Suppose a class needs a value for pi calculated to particular tolerances. The class could define its own version of the built-in pi function for use within the class. This approach maintains the encapsulation of the classs internal workings, but does not require an instance of the class to return a value.
Defining a Static Method

To define a method as static, set the methods block Static attribute to true. For example:
classdef MyClass ... methods(Static) function p = pi(tol) [n d] = rat(pi,tol); p = n/d; end end end
7-25
Update Graphs Using Events and Listeners on page 9-35 provides an example that uses a static method to create a set of objects representing graphs.
Calling Static Methods

Invoke static methods using the name of the class followed by dot (.), then the name of the method:
classname.staticMethodName(args,...)
Calling the pi method of MyClass in the previous section would require this statement:
value = MyClass.pi(.001);
You can also invoke static methods using an instance of the class, like any method:
obj = MyClass; value = obj.pi(.001);
createViews static method provides an example of a static method.
Inheriting Static Methods

Subclasses can redefine static methods unless the methods Sealed attribute is also set to true in the superclass.
7-26
Overloading Functions for Your Class
Overloading Functions for Your Class

In this section... Overloading MATLAB Functions on page 7-27 Rules for Naming to Avoid Conflicts on page 7-28
Overloading MATLAB Functions

Class methods can provide implementations of MATLAB functions that operate only on instances of the class. This is possible because MATLAB software can always identify which class an object belongs to. MATLAB uses the dominant argument to determine which version of a function to call. If the dominant argument is an object, then MATLAB calls the method defined by the objects class, if there is one. In cases where a class defines a function with the same name as a global function, the classs implementation of the function is said to overload the original global implementation. Note MATLAB does not support overloading functions using different signatures for the same function name.
Implementing MATLAB Operators

Classes designed to implement new MATLAB data types typically define certain operators, such as addition, subtraction, equality, and so on. For example, standard MATLAB addition (+) cannot add two polynomials because this operation is not defined by simple addition. However, a polynomial class can define its own plus method that the MATLAB language calls to perform addition of polynomial objects when you use the + symbol:
p1 + p2
Implementing Operators for Your Class on page 15-31 provides information on methods to overload.
7-27
Defining Arithmetic Operators for DocPolynom on page 17-13 provides examples.
Rules for Naming to Avoid Conflicts

The names of methods, properties, and events are scoped to the class. Therefore, you should adhere to the following rules to avoid naming conflicts: You can reuse names that you have used in unrelated classes. You can reuse names in subclasses if the member does not have public or protected access. These names then refer to entirely different methods, properties, and events without affecting the superclass definitions Within a class, all names exist in the same name space and must be unique. A class cannot define two methods with the same name and a class cannot define a local function with the same name as a method. The name of a static method is considered without its class prefix. Thus, a static method name without its class prefix cannot match the name of any other method.
7-28
Object Precedence in Expressions Using Operators
Object Precedence in Expressions Using Operators

Establishing an object precedence enables the MATLAB runtime to determine which of possibly many versions of an operator or function to call in a given situation. For example, consider the expression
objectA + objectB
Ordinarily, objects have equal precedence and the method associated with the left-most object is called. However, there are two exceptions: User-defined classes have precedence over MATLAB built-in classes. User-defined classes can specify their relative precedence with respect to other user-defined classes using the InferiorClasses attribute. In A Polynomial Class on page 17-2, the polynom class defines a plus method that enables the addition of DocPolynom objects. Given the object p:
p = DocPolynom([1 0 -2 -5]) p = x^3-2*x-5
the expression:
1 + p ans = x^3-2*x-4
calls the DocPolynom plus method (which converts the double, 1, to a DocPolynom object and then implements the addition of two polynomials). The user-defined DocPolynom class has precedence over the builtin double class.
Specifying Precedence of User-Defined Classes

You can specify the relative precedence of user-defined classes by listing inferior classes using a class attribute. The InferiorClasses property places a class below other classes in the precedence hierarchy. Define the InferiorClasses property in the classdef statement:
7-29
classdef (InferiorClasses = {?class1,?class2}) myClass
This attribute establishes a relative priority of the class being defined with the order of the classes listed.
Location in the Hierarchy

If objectA is above objectB in the precedence hierarchy, then the expression
objectA + objectB
calls @classA/plus.m. Conversely, if objectB is above objectA in the precedence hierarchy, then the MATLAB runtime calls @classB/plus.m. See Rules for Naming to Avoid Conflicts on page 7-28 for related information.
7-30
Class Methods for Graphics Callbacks

In this section... Callback Arguments on page 7-31 General Syntax for Callbacks on page 7-31 Object Scope and Anonymous Functions on page 7-32 Example Class Method as a Slider Callback on page 7-33
Callback Arguments
You can use class methods as callbacks for Handle Graphics objects by specifying the callback as an anonymous function. Anonymous functions enable you to pass the arguments required by methods (i.e., the first argument is a class object) and graphics object callbacks (i.e., the event source and the event data), as well as any other arguments you want to pass to the function. The following links provide general information on graphics object callbacks and anonymous functions.
Background Information
Function Handle Callbacks Information on graphics object callbacks Anonymous Functions Information about using anonymous functions
General Syntax for Callbacks

The basic syntax for a function handle that you assign to the graphic objects Callback property includes the object as the first argument:
@(src,event)method_name(object,src,event,additional_arg,...)
You must define the callback method with the following signature:
method_name(object,src,event)
7-31
Object Scope and Anonymous Functions

Anonymous functions take a snapshot of the argument values when you define the function handle. You must, therefore, consider this scoping when assigning the Callback property. The following two sections provide examples.
Using Value Classes

Consider the following snippet of a value class definition:
classdef SeaLevelAdjuster properties Slider end methods function seal = SeaLevelAdjuster
...
seal.Slider = uicontrol('Style','slider'); set(seal.Slider,'Callback',@(src,event)slider_cb(seal,src,event)) end end end
This class assigns the Callback property in a separate set statement so that the value objects (seal) Slider property has been defined when you create the function handle. Otherwise, Handle Graphics freezes seal before the uicontrols handle is assigned to the Slider property.
Using Handle Classes

The difference in behavior between a handle object and a value object is important in this case. If you defined the class as a handle class, the object is a reference to the underlying data. Therefore, when the MATLAB runtime resolves the function handle, the contents of the object reflects assignments made after the function handle is defined:
classdef SeaLevelAdjuster < handle
...
properties Slider end
7-32
methods function seal = SeaLevelAdjuster

...
seal.Slider = uicontrol('Style','slider',... 'Callback',@(src,event)slider_cb(seal,src,event)); end end end
Example Class Method as a Slider Callback

This example defines a slider that varies the color limits of an indexed image to give the illusion of varying the sea level.
Displaying the Class Files

Open the SeaLevelAdjuster class definition file in the MATLAB editor. To use the class, create a folder named @SeaLevelAdjuster and save SeaLevelAdjuster.m to this folder. The parent folder of @SeaLevelAdjuster must be on the MATLAB path.
Class Properties
The class defines properties to store graphics object handles and the calculated color limits:
classdef SeaLevelAdjuster < handle properties Figure = []; Axes = []; Image = []; CLimit = []; Slider = []; end end
7-33
Class Constructor
The class constructor creates the graphics objects and assigns the slider callback (last line in code snippet):
methods function seal = SeaLevelAdjuster(x,map) seal.Figure = figure('Colormap',map,... 'Resize','off',... 'Position',[100 100 560 580]); seal.Axes = axes('DataAspectRatio',[1 1 1],... 'XLimMode','manual',... 'YLimMode','manual',... 'Parent',seal.Figure); seal.Image = image(x,'CDataMapping','scaled','Parent',seal.Axes); seal.CLimit = get(seal.Axes,'CLim'); seal.Slider = uicontrol('Style','slider',... 'Parent',seal.Figure,... 'Max',seal.CLimit(2),... 'Min',seal.CLimit(1)-1,... 'Value',seal.CLimit(1),... 'Units','normalized',... 'Position',[.9286 .1724 .0357 .6897],... 'SliderStep',[0.002 0.005 ],... 'Callback',@(src,event)slider_cb(seal)); end % SeaLevelAdjuster end % methods
The callback function for the slider is defined to accept the three required arguments a class instance, the handle of the event source, and the event data:
methods function slider_cb(seal) min_val = get(seal.Slider,'Value'); max_val = max(max(get(seal.Image,'CData'))); set(seal.Axes,'CLim',[min_val max_val]) drawnow end % slider_cb end % methods
7-34
Using the SeaLevelAdjuster Class

The class is designed to be used with the cape image that is included with the MATLAB product. To obtain the image data, use the load command:
load cape
After loading the data, create a SeaLevelAdjuster object for the image:
seal = SeaLevelAdjuster(X,map);
Move the slider to change the apparent sea level and visualize what would happen to Cape Cod if the sea level were to rise.
7-35
50
100
150
200
250
300
350 50 100 150 200 250 300 350
7-36
8
Object Arrays
Creating Object Arrays on page 8-2 Concatenating Objects of Different Classes on page 8-13
Object Arrays
Creating Object Arrays

In this section... Basic Knowledge on page 8-2 Building Arrays in the Constructor on page 8-2 Initializing Arrays of Value Objects on page 8-3 Initial Value of Object Properties on page 8-5 Creating Empty Arrays on page 8-6 Initializing Arrays of Handle Objects on page 8-7 Referencing Property Values in Object Arrays on page 8-9 Object Arrays with Dynamic Properties on page 8-10
Basic Knowledge
The material presented in this section builds on an understanding of the information presented in the following sections. Class Definitions Class Syntax Fundamentals Class Constructor Methods on page 7-16 Working with Arrays Matrix Indexing Empty Matrices, Scalars, and Vectors Multidimensional Arrays
Building Arrays in the Constructor

A constructor method can create an object array by building the array and returning it as the output argument.
8-2
For example, the following DocArrayExample class creates an object array the same size as the input array and initializes the Value property of each object to the corresponding input array value.
classdef DocArrayExample properties Value end methods function obj = DocArrayExample(F) if nargin ~= 0 % Allow nargin == 0 syntax m = size(F,1); n = size(F,2); obj(m,n) = DocArrayExample; % Preallocate object array for i = 1:m for j = 1:n obj(i,j).Value = F(i,j); end end end end end end
To preallocate the object array, assign the last element of the array first. MATLAB fills the first to penultimate array elements with default DocArrayExample objects. After preallocating the array, assign each object Value property to the corresponding value in the input array F. For example:
F = magic(5); % Create 5-by-5 array of magic square numbers A = DocArrayExample(F); % Create 5-by-5 array of objects
Initializing Arrays of Value Objects

During the creation of object arrays, MATLAB might need to call the class constructor with no arguments, even if the constructor does not build an object array. For example, suppose you define the following class:
classdef SimpleClass
8-3
Object Arrays
properties Value end methods function obj = SimpleClass(v) obj.Value = v; end end end
Now execute the following statement (which is a valid MATLAB statement):

a(1,7) = SimpleClass(7) Error using SimpleClass>SimpleClass.SimpleClass Not enough input arguments.
This error occurs because MATLAB calls the constructor with no arguments to initialize elements 1 through 6 in the array (that is, a(1,1:6)). Therefore, you must ensure the constructor supports the no input argument syntax. A simple solution is to test nargin and let the case when nargin == 0 execute no code, but not error:
classdef SimpleClass properties Value end methods function obj = SimpleClass(v) if nargin > 0 obj.Value = v; end end end end
Using the revised class definition, the previous array assignment statement executes without error:
a(1,7) = SimpleClass(7) a =
8-4
1x7 SimpleClass array with properties: Value
The object assigned to array element a(1,7) uses the input argument passed to the constructor as the value assigned to the property:
a(1,7) ans = SimpleClass with properties: Value: 7
However, MATLAB created the objects contained in elements a(1,1:6) with no input argument and initialized the value of the Value property to empty []. For example:
a(1,1) ans = SimpleClass with properties: Value: []
MATLAB calls the SimpleClass constructor once and copies the returned object to each element of the array.
Initial Value of Object Properties

When MATLAB calls a constructor with no arguments to initialize an object array, one of the following assignments occurs: If property definitions specify default values, MATLAB assigns these values. If the constructor assigns values in the absence of input arguments, MATLAB assigns these values. If neither of the above situations apply, MATLAB assigns the value of empty double (i.e., []) to the property.
8-5
Object Arrays
Creating Empty Arrays

Empty arrays have no elements, but are of a certain class. All nonabstract classes have a static method named empty that creates an empty array of the same class. The empty method enables you to specify the dimensions of the output array. However, at least one of the dimensions must be 0. For example:
ary = SimpleClass.empty(5,0);
creates a 5by0 empty array of class SimpleClass. Calling empty with no arguments returns a 0by0 empty array.
Assigning Values to an Empty Array

An empty object defines the class of an array. However, to assign nonempty objects to an empty array, MATLAB must call the class constructor to create default instances of the class for every other array element. Once you assign a nonempty object to an array, all array elements must be nonempty objects. Note A class constructor should never return empty objects by default. For example, using the SimpleClass defined in the Initializing Arrays of Value Objects on page 8-3 section, create an empty array:
ary = SimpleClass.empty(5,0); class(ary) ans = SimpleClass
The array ary is an array of class SimpleClass. However, it is an empty array:

ary(1) Index exceeds matrix dimensions.
If you make an assignment to a property value, MATLAB calls the SimpleClass constructor to grow the array to the require size:
8-6
ary(5).Value = 7; ary(5).Value ans = 7 ary(1).Value ans = []
In this case, MATLAB populates array elements one through five with SimpleClass objects created by calling the class constructor with no arguments. Then MATLAB assigns the property value 7 to the object at ary(5).
Initializing Arrays of Handle Objects

When MATLAB expands an array, it calls the class constructor once, and then creates unique handles for each element in the array. MATLAB copies the property values from the constructed object without calling the constructor for each additional element. The following class illustrates this behavior.
classdef InitArray < handle properties RandNumb end methods function obj = InitArray obj.RandNumb = randi(100); end end end
8-7
Object Arrays
The property RandNumb contains a random number that is assigned from the InitArray class constructor. The next section uses the InitArray class to show when MATLAB calls the class constructor when expanding an array.
Initializing a Handle Object Array

Consider what happens when MATLAB initialize an array by first assigning to the last element in the array. (The last element is the one with the highest index values). For example, suppose the value of the RandNumb property of the InitArray object assigned to the element A(4,5) is 59:
A(4,5) = InitArray; A(4,5).RandNumb ans = 59
The element in the index location A(4,5) is an instance of the InitArray class. Element A(1,1) is also an instance of the InitArray class, but its RandNumb property is set to a different random number. The difference is caused by the fact that MATLAB called the class constructor to create a single object, which MATLAB then copied to all the remaining array elements. Calling the constructor resulted in another call to the randi function, which returns a new random number:
A(1,1).RandNumb ans = 91
MATLAB copies this second instance to all remaining array elements:

A(2,2).RandNumb ans = 91 A(2,3).RandNumb
8-8
ans = 91
When initializing an object array, MATLAB assigns a copy of a single object to the empty elements in the array. However, MATLAB gives each object a unique handle so that later you can assign different property values to each object. This means that the objects are not equivalent:
A(1,1) == A(2,2) ans = 0
Therefore, the creation of an array with a statement such as:

A(4,5) = InitArray;
results in two calls to the class constructor. The first creates the object for array element A(4,5). The second creates a default object (no arguments passed to the constructor) that MATLAB copies to all remaining empty array elements. See Indexing Multidimensional Arrays and Reshaping Multidimensional Arrays for information on array manipulation. See Initializing Properties to Unique Values on page 3-12 for information on assigning values to properties. See Indexed Reference and Assignment on page 15-9 for information on implementing subsasgn methods for your class.
Referencing Property Values in Object Arrays

You can reference all values of the same property in an object array using the syntax:
objarray.PropName
For example, given the ObjArray class:
8-9
Object Arrays
classdef ObjArray properties RegProp end methods function obj = ObjArray % Assign property a random integer obj.RegProp = randi(100); end end end
Create an array of ObjArray objects and assign all values of the RegProp property to the propvalues cell array:
for k = 1:5 a(k) = ObjArray; end propvalues = {a.RegProp} propvalues = [96] [49] [81] [15] [43]
Object Arrays with Dynamic Properties

See Dynamic Properties Adding Properties to an Instance on page 6-27 for information about classes that can define dynamic properties. You cannot reference all the dynamic properties in an object array using a single statement, as shown in the previous section for ordinary properties. For example, suppose the ObjArray class subclasses the dynamicprops class, which enables you to add properties to instances of the ObjArray class.
classdef ObjArray < dynamicprops properties RegProp end methods function obj = ObjArray % Assign property a random integer
8-10
obj.RegProp = randi(100); end end end
Create an object array and add dynamic properties to each member of the array:
% Define elements 1 and 2 as ObjArray objects a(1) = ObjArray; a(2) = ObjArray; % Add dynamic properties to each object and assign a value a(1).addprop('DynoProp'); a(1).DynoProp = 1; a(2).addprop('DynoProp'); a(2).DynoProp = 2;
You can get the values of the ordinary properties, as with any array:
a.RegProp ans = 4 ans = 85
MATLAB returns an error if you try to access the dynamic properties of all array elements using this syntax.
a.DynoProp No appropriate method, property, or field DynoProp for class ObjArray.
You must refer to each object individually to access dynamic property values:
a(1).DynoProp ans =
8-11
Object Arrays
1 a(2).DynoProp ans = 2
8-12
Concatenating Objects of Different Classes

In this section... Basic Knowledge on page 8-13 MATLAB Concatenation Rules on page 8-13 Concatenating Objects on page 8-14 Converting to the Dominant Class on page 8-14 Implementing Converter Methods on page 8-17
Basic Knowledge
The material presented in this section builds on an understanding of the information presented in the following sections. Class Precedence on page 4-18 Class Attributes on page 4-6 Creating Object Arrays on page 8-2
MATLAB Concatenation Rules

MATLAB follows these rules for concatenating objects: MATLAB always converts all objects to the dominant class. User-defined classes take precedence over built-in classes like double. If there is no defined dominance relationship between any two objects, then the left-most object dominates (see Class Precedence on page 4-18). MATLAB does not convert objects to a common superclass unless those objects derive from a member of a heterogeneous hierarchy. See matlab.mixin.Heterogeneous for more information. The following sections describe these rules in more detail. See Valid Combinations of Unlike Classes for related information.
8-13
Object Arrays
Concatenating Objects
Concatenation combines objects into arrays:
ary = [obj1,obj2,obj3,...,objn]; % size of ary is 1-by-n ary = [obj1;obj2;obj3;...;objn]; % size of ary is n-by-1
The class of the resulting array, ary, is the same as the class of the objects being concatenated. Concatenating unlike objects is possible if MATLAB can convert objects to the dominant class. MATLAB attempts to convert unlike objects by: Calling the inferior objects converter method, if one exists (see Implementing Converter Methods on page 8-17 for an example). Passing an inferior object to the dominant class constructor to create an object of the dominant class. If conversion of the inferior objects is successful, MATLAB returns an array that is of the dominant class. If conversion is not possible, MATLAB returns an error.
Modifying Default Concatenation

Classes can control how concatenation of its instances works by overloading horzcat, vertcat, cat. See Redefining Concatenation for Your Class on page 15-6 for more information.
Converting to the Dominant Class

MATLAB first attempts to find converter methods for objects of the inferior classes. If your class design requires object conversion, implement methods for this purpose. See Converting Objects to Another Class on page 15-7 for general information on converter methods.
Calling the Dominant-Class Constructor

When MATLAB calls the dominant class constructor to convert an object of an inferior class to the dominant class, the inferior object is passed to the constructor as an argument. If the class design enables the dominant class constructor to accept objects of inferior classes as input arguments, then concatenation is possible without implementing a separate converter method.
8-14
In cases where the constructor simply assigns this argument to a property, the result is an object of the dominant class with an object of an inferior class stored in a property. If this is not a desired result, then ensure that class constructors include adequate error checking. For example, consider the class ColorClass and two subclasses, RGBColor and HSVColor:
classdef ColorClass properties Color end end
The class RGBColor inherits the Color property from ColorClass. RGBColor stores a color value defined as a three-element vector of red, green, and blue (RGB) values. The constructor does not restrict the value of the input argument. It assigns this value directly to the Color property.
classdef RGBColor < ColorClass % Class to contain RGB color specification methods function obj = RGBColor(rgb) if nargin > 0 obj.Color = rgb; end end end end
The class HSVColor also inherits the Color property from ColorClass. HSVColor stores a color value defined as a three-element vector of hue, saturation, brightness value (HSV) values.
classdef HSVColor < ColorClass % Class to contain HSV color specification methods function obj = HSVColor(hsv) if nargin > 0 obj.Color = hsv; end
8-15
Object Arrays
end end end
Create an instance of each class and concatenate them into an array. The RGBColor object is dominant because it is the left most object and neither class defines a dominance relationship:
crgb = RGBColor([1 0 0]); chsv = HSVColor([0 1 1]); ary = [crgb,chsv]; class(ary) ans = RGBColor
MATLAB can combine these different objects into an array because it can pass the inferior object of class HSVColor to the constructor of the dominant class. However, notice that the Color property of the second RGBColor object in the array actually contains an HSVColor object, not an RGB color specification:
ary(2).Color ans = HSVColor with properties: Color: [0 1 1]
Avoid this undesirable behavior by: Implementing converter methods Performing argument checking in class constructors before assigning values to properties The next section shows updates to these classes.
8-16
Implementing Converter Methods

Here is the ColorClass class with converter methods for RGBColor and HSVColor objects:
classdef ColorClass properties Color end methods function rgbObj = RGBColor(obj) % Convert HSVColor object to RGBColor object if strcmp(class(obj),'HSVColor') rgbObj = RGBColor(hsv2rgb(obj.Color)); end end function hsvObj = HSVColor(obj) % Convert RGBColor object to HSVColor object if strcmp(class(obj),'RGBColor') hsvObj = HSVColor(rgb2hsv(obj.Color)); end end end end
Create an array of RGBColor and HSVColor objects with the revised superclass:
crgb = RGBColor([1 0 0]); chsv = HSVColor([0 1 1]); ary = [crgb,chsv]; class(ary) ans = RGBColor
MATLAB calls the converter method for the HSVColor object, which it inherits from the superclass. The second array element is now an RGBColor object with an RGB color specification assigned to the Color property:
ary(2)
8-17
Object Arrays
ans = RGBColor with properties: Color: [1 0 0] ary(2).Color ans = 1 0 0
If the left-most object is of class HSVColor, the array ary is also of class HSVColor, and MATLAB converts the Color property data to HSV color specification.
ary = [chsv crgb] ary = 1x2 HSVColor Properties: Color ary(2).Color ans = 0 1 1
Defining a converter method in the superclass and adding better argument checking in the subclass constructors produces more predicable results. Here is the RGBColor class constructor with argument checking:
classdef RGBColor < ColorClass2 methods function obj = RGBColor(rgb) if nargin == 0 rgb = [0 0 0]; else if ~(strcmp(class(rgb),'double')... && size(rgb,2) == 3 && max(rgb) <= 1 && min(rgb) >= 0)
8-18
error('Specify color as RGB values') end end obj.Color = rgb; end end end
Your applications might require additional error checking and other coding techniques. The classes in these examples are designed only to demonstrate concepts. See Class Constructor Methods on page 7-16 for more information on writing class constructors. See Hierarchies of Classes Concepts on page 10-2 for more information on inheritance.
8-19
Object Arrays
8-20
9
Events Sending and Responding to Messages
Learning to Use Events and Listeners on page 9-2 Create a Property Set Listener on page 9-8 Events and Listeners Concepts on page 9-11 Event Attributes on page 9-16 Events and Listeners Syntax and Techniques on page 9-18 Listen for Changes to Property Values on page 9-27 Update Graphs Using Events and Listeners on page 9-35
Learning to Use Events and Listeners

In this section... Why Use Events and Listeners on page 9-2 What You Need to Know on page 9-2 Customizing Event Data on page 9-3 Observe Property Changes on page 9-6
Why Use Events and Listeners

Events are notices that objects broadcast in response to something that happens, such as a property value changing or a user interaction with an application program. Listeners execute functions when notified that the event of interest occurs. Use events to communicate things that happen to objects, and respond to these events by executing the listeners callback function. See Events and Listeners Concepts on page 9-11 for a more thorough discussion of the MATLAB event model.
What You Need to Know

The following sections provide simple examples that show the basic techniques for using events and listeners. Subsequent sections provide more detailed descriptions and more complex examples.
Events and Listeners Basics

When using events and listeners: Only handle classes can define events and listeners (See Naming Events on page 9-18 for syntax). Call the handle notify method to trigger the event (See Triggering Events on page 9-18, and Defining and Triggering an Event on page 9-4, for examples). The event notification broadcasts the named event to all listeners registered for this event. Use the handle addlistener method to associate a listener with an object that will be the source of the event (Listening to Events on page 9-19,
9-2
Creating a Listener for the Overflow Event on page 9-5, and Property Event and Listener Classes on page 9-29). When adding a listener, pass a function handle for the listener callback function using a syntax such as the following:
addlistener(eventObject,'EventName',@functionName) for an
ordinary function.
addlistener(eventObject,'EventName',@Obj.methodName) for a
method of Obj.
addlistener(eventObject,'EventName',@ClassName.methodName)
for a static method of the class ClassName.
Listener callback functions must define at least two input arguments the event source object handle and the event data (See Defining Listener Callback Functions on page 9-24 for more information). You can modify the data passed to each listener callback by subclassing the event.EventData class (See Defining Event-Specific Data on page 9-21) and Defining the Event Data on page 9-5 for more information).
Customizing Event Data

Suppose you want to create a listener callback function that has access to specific information when the event occurs. This example shows how to do this by creating custom event data. Events provide information to listener callback functions by passing an event data argument to the specified function. By default, MATLAB passes an event.EventData object to the listener callback. This object has two properties: EventName Name of the event triggered by this object. Source Handle of the object triggering the event. Provide additional information to the listener callback by subclassing the event.EventData class. Define properties in the subclass to contain the additional data. Define a constructor that accepts the additional data as arguments.
9-3
Use the subclass constructor as an argument to the notify method to trigger the event. The Defining the Event Data on page 9-5 section shows an implementation of this subclass. See Defining Event-Specific Data on page 9-21 for another example that subclasses event.EventData.
Defining and Triggering an Event

The SimpleEventClass defines a property set method (see Property Set Methods on page 6-16) from which it triggers an event if the property is set to a value exceeding a certain limit. The property set method performs these operations: Saves the original property value Sets the property to the specified value If the specified value is greater than 10, the set method triggers an Overflow event Passes the original property value, as well as other event data, in a SpecialEventDataClass object to the notify method (see Defining the Event Data on page 9-5 )
classdef SimpleEventClass < handle % Must be a subclass of handle properties Prop1 = 0; end events Overflow end methods function set.Prop1(obj,value) orgvalue = obj.Prop1; obj.Prop1 = value; if (obj.Prop1 > 10) % Trigger the event using custom event data notify(obj,'Overflow',SpecialEventDataClass(orgvalue));
9-4
end end end end
Defining the Event Data

Event data is always contained in an event.EventData object. The SpecialEventDataClass adds the original property value to the event data by subclassing event.EventData:
classdef SpecialEventDataClass < event.EventData properties OrgValue = 0; end methods function eventData = SpecialEventDataClass(value) eventData.OrgValue = value; end end end
Creating a Listener for the Overflow Event

To listen for the Overflow event, attach a listener to an instance of the SimpleEventClass class. Use the addlistener method to create the listener. You also need to define a callback function for the listener to execute when the event is triggered. The function setupSEC instantiates the SimpleEventClass class and adds a listener to the object. In this example, the listener callback function displays information that is contained in the eventData argument (which is a SpecialEventDataClass object).
function sec = setupSEC % Create an object and attach the listener sec = SimpleEventClass; addlistener(sec,'Overflow',@overflowHandler) % Define the listener callback function function overflowHandler(eventSrc,eventData) disp('The value of Prop1 is overflowing!')
9-5
disp(['It''s value was: ' num2str(eventData.OrgValue)]) disp(['It''s current value is: ' num2str(eventSrc.Prop1)]) end end
Create the SimpleEventClass object and add the listener:

>> sec = setupSEC; >> sec.Prop1 = 5; >> sec.Prop1 = 15; % listener triggers callback The value of Prop1 is overflowing! It's value was: 5 It's current value is: 15
Observe Property Changes

This example shows how to listen for changes to a property value. This examples uses: PostSet event predefined by MATLAB SetObservable property attribute to enable triggering the property PostSet event. addlistener handle class method to create the listener
classdef PropLis < handle % Define a property that is SetObservable properties (SetObservable) ObservedProp = 1; end methods function attachListener(obj) %Attach a listener to a PropListener object addlistener(obj,'ObservedProp','PostSet',@PropLis.propChange); end end methods (Static) function propChange(metaProp,eventData) % Callback for PostSet event % Inputs: meta.property object, event.PropertyEvent h = eventData.AffectedObject;
9-6
propName = metaProp.Name; disp(['The ',propName,' property has changed.']) disp(['The new value is: ',num2str(h.ObservedProp)]) disp(['It''s defaul value is: ',num2str(metaProp.DefaultValue)]) end end end
The PropLis class uses an ordinary method (attachListener) to add the listener for the ObservedProp property. If the PropLis classed defined a constructor, it could contain the call to addlistener. The listener callback is a static method (propChange) to which MATLAB passes a meta.property object for ObservedProp, and an event.PropertyEvent object. These arguments provide information about the property and the event. Use the PropLis class by creating an instance and calling its attachListener method:
plObj = PropLis; plObj.ObservedProp ans = 1 plObj.attachListener plObj.ObservedProp = 2; The ObservedProp property has changed. The new value is: 2 It's defaul value is: 1
See Creating Property Listeners on page 9-27 for more information on property listeners.
9-7
Create a Property Set Listener

This example shows how to define a listener for a property set event. This means the listener callback triggers when the value of a specific property changes. The class defined for this example uses a method for a push button callback and a static method for the listener callback. When the push button callback changes the value of a property, the listener executes its callback on the PreSet event. This example defines a class (PushButton) with these design elements: ResultNumber Observable property uicontrol pushbutton Push button object used to generate a new graph when its callback executes A listener that responds to a change in the observable property The PushButton Class Design The PushButton class creates figure, uicontrol, axes graphics objects and a listener object in the class constructor. The push buttons callback is a class method (named pressed). When the push button is activated, the following sequence occurs:
1 MATLAB executes the pressed method, which graphs a new set of data
and increments the ResultNumber property.

2 Attempting to set the value of the ResultNumber property triggers the
PreSet event, which executes the listener callback before setting the property value.
3 The listener callback uses the event data to obtain the handle of the
callback object (an instance of the PushButton class), which then provides the handle of the axes object that is stored in its AxHandle property.
4 The listener callback updates the axes Title property, after the callback
completes execution, MATLAB sets the ResultsNumber property to its new value.
9-8
Create a Property Set Listener
classdef PushButton < handle properties (SetObservable) % Enable listener to observe PreSet event ResultNumber = 1; end properties AxHandle end methods function buttonObj = PushButton myFig = figure; buttonObj.AxHandle = axes('Parent',myFig); uicontrol('Parent',myFig,... 'Style','pushbutton',... 'String','Plot Data',... 'Callback',@(src,evnt)pressed(buttonObj)); addlistener(buttonObj,'ResultNumber','PreSet',... @PushButton.updateTitle); end end methods function pressed(obj) % Push button callback % Perform plotting operations % Specify target axes for output of plot function scatter(obj.AxHandle,randn(1,20),randn(1,20),'p') % Update observed property % which triggers PreSet event obj.ResultNumber = obj.ResultNumber + 1; end end methods (Static) function updateTitle(~,eventData) % Listener callback - updates axes title h = eventData.AffectedObject; % Get handle to title text object % And set String property set(get(h.AxHandle,'Title'),'String',['Result Number: ',... num2str(h.ResultNumber)]) end
9-9
end end
The scatter graph looks similar to this after three push-button clicks:
See Listen for Changes to Property Values on page 9-27 for more on property events.
9-10
Events and Listeners Concepts

In this section... The Event Model on page 9-11 Default Event Data on page 9-13 Events Only in Handle Classes on page 9-13 Property-Set and Query Events on page 9-14 Listeners on page 9-15
The Event Model

Events represent changes or actions that occur within class instances. For example, Modification of class data Execution of a method Querying or setting a property value Destruction of an object Basically, any activity that you can detect programmatically can generate an event and communicate information to other objects. MATLAB classes define a process that communicates the occurrence of events to other objects that need to respond to the events. The event model works this way: A handle class declares a name used to represent an event. Naming Events on page 9-18 After creating an instance of the event-declaring class, you can attach listener objects to it. Ways to Create Listeners on page 9-22 A call to a class method broadcasts a notice of the event to listeners. The class user determines when to trigger the event. Triggering Events on page 9-18
9-11
Listeners execute a callback function when notified that the event has occurred. Defining Listener Callback Functions on page 9-24 You can bind listeners to the lifecycle of the object that defines the event, or limit listeners to the existence and scope of the listener object. Ways to Create Listeners on page 9-22 The following diagram illustrates the event model.
1. The withdraw method is called. if AccountBalance <= 0 notify(obj,InsufficientFunds); end
BankAccount Properties AccountNumber AccountBalance Methods deposit withdraw Events InsufcientFunds
2. The notify method triggers an event, and a message is broadcast.
InsufcientFunds 3. Listeners awaiting message execute their callbacks. (The broadcasting object does not necessarily know who is listening.) Listener1 Properties EventName = InsufcientFunds FunctionHandle = @Callback1
InsufcientFunds
Listener2 Properties EventName = InsufcientFunds FunctionHandle = @Callback2
9-12
Default Event Data

Events provide information to listener callbacks by passing an event data argument to the callback function. By default, MATLAB passes an event.EventData object to the listener callback. This object has two properties: EventName The event name as defined in the class event block Source The object that is the source of the event MATLAB passes the source object to the listener callback in the required event data argument. This enables you to access any of the objects public properties from within your listener callback function.
Customizing Event Data

You can create a subclass of the event.EventData class to provide additional information to listener callback functions. The subclass would define properties to contain the additional data and provide a method to construct the derived event data object so it can be passed to the notify method. Defining Event-Specific Data on page 9-21 provides an example showing how to customize this data.
Events Only in Handle Classes

You can define events only in handle classes. This restriction exists because a value class is visible only in a single MATLAB workspace so no callback or listener can have access to the object that triggered the event. The callback could have access to a copy of the object. However, accessing a copy is not generally useful because the callback cannot access the current state of the object that triggered the event or effect any changes in that object. Comparing Handle and Value Classes on page 5-2 provides general information on handle classes. Events and Listeners Syntax and Techniques on page 9-18 shows the syntax for defining a handle class and events.
9-13
Property-Set and Query Events

There are four predefined events related to properties: PreSet Triggered just before the property value is set, before calling its set access method PostSet Triggered just after the property value is set PreGet Triggered just before a property value query is serviced, before calling its get access method PostGet Triggered just after returning the property value to the query These events are predefined and do not need to be listed in the class events block. When a property event occurs, the callback is passed an event.PropertyEvent object. This object has three properties: EventName The name of the event described by this data object Source The source object whose class defines the event described by the data object AffectedObject The object whose property is the source for this event (that is, AffectedObject contains the object whose property was either accessed or modified). You can define your own property-change event data by subclassing the event.EventData class. Note that the event.PropertyEvent class is a sealed subclass of event.EventData. See Listen for Changes to Property Values on page 9-27 for a description of the process for creating property listeners. See The PostSet Event Listener on page 9-48 for an example. See Property Access Methods on page 6-14 for information on methods that control access to property values.
9-14
Listeners
Listeners encapsulate the response to an event. Listener objects belong to the event.listener class, which is a handle class that defines the following properties: Source Handle or array of handles of the object that generated the event EventName Name of the event Callback Function to execute when an enabled listener receives event notification Enabled Callback function executes only when Enabled is true. See Enabling and Disabling the Listeners on page 9-51 for an example. Recursive Allow listener to cause the same event that triggered the execution of the callback
Recursive is true by default. It is possible to create a situation where infinite recursion reaches the recursion limit and eventually triggers an error. If you set Recursive to false, the listener cannot execute recursively if the callback triggers its own event.
Ways to Create Listeners on page 9-22 provides more specific information.
9-15
Event Attributes
Table of Event Attributes
The following table lists the attributes you can set for events. To specify a value for an attribute, assign the attribute value on the same line as the event key word. For example, all the events defined in the following events block have private ListenAccess and NotifyAccess attributes.
events (ListenAccess = 'private', NotifyAccess = 'private') anEvent anotherEvent end
To define other events in the same class definition that have different attribute settings, create another events block. Attribute Name
Hidden
Class
logical Default = false
Description If true, event does not appear in list of events returned by events function (or other event listing functions or viewers). Determines where you can create listeners for the event. public Unrestricted access protected Access from methods in class or subclasses private Access by class methods only (not from subclasses) List classes that have listen access to this event. Specify classes as meta.class objects in the form:
ListenAccess enumeration, default = public

meta.class
objects
9-16
Event Attributes
Class
Description See Controlling Access to Class Members on page 10-24
NotifyAccess enumeration, default = public
Determines where code can trigger the event public Any code can trigger event protected Can trigger event from methods in class or derived classes private Can trigger event by class methods only (not from derived classes) List classes that have notify access to this event. Specify classes as meta.class objects in the form:

meta.class
objects
See Controlling Access to Class Members on page 10-24
9-17
Events and Listeners Syntax and Techniques

In this section... Naming Events on page 9-18 Triggering Events on page 9-18 Listening to Events on page 9-19 Defining Event-Specific Data on page 9-21 Ways to Create Listeners on page 9-22 Defining Listener Callback Functions on page 9-24 Callback Execution on page 9-26
Naming Events
Define an event by declaring an event name inside an events block, typically in the class that generates the event. For example, the following class creates an event called ToggledState, which might be triggered whenever a toggle buttons state changes.
classdef ToggleButton < handle properties State = false end events ToggledState end end
Triggering Events
At this point, the ToggleButton class has defined a name that it associates with the toggle button state changestoggling on and toggling off. However, a class method controls the actual firing of the events. To accomplish this, the ToggleButton class adds a method to trigger the event:
classdef ToggleButton < handle properties State = false
9-18
end events ToggledState end methods ... function OnStateChange(obj,newState) % Call this method to check for state change if newState ~= obj.State obj.State = newState; notify(obj,'ToggledState'); % Broadcast notice of event end end end end
The OnStateChange method calls notify to trigger the event, using the handle of the ToggleButton object that owns the event and the string name of the event.
Listening to Events
Once the call to notify triggers an event, MATLAB broadcasts a message to all registered listeners. To register a listener for a specific event, use the addlistener handle class method. For example, the following class defines objects that listen for the ToggleState event defined in the class ToggleButton.
classdef RespondToToggle < handle methods function obj = RespondToToggle(toggle_button_obj) addlistener(toggle_button_obj,'ToggledState',@RespondToToggle.handleEvnt); end end methods (Static) function handleEvnt(src,evtdata) if src.State disp('ToggledState is true') else disp('ToggledState is false') % Respond to false ToggleState here % Respond to true ToggleState here
9-19
end end end end
The class RespondToToggle adds the listener from within its constructor. The class defines the callback (handleEvnt) as a static method that accepts the two standard arguments: src the handle of the object triggering the event (i.e., a ToggleButton object) evtdata an event.EventData object The listener executes the callback when the specific ToggleButton object executes the notify method, which it inherits from the handle class. For example, create instances of both classes:
tb = ToggleButton; rtt = RespondToToggle(tb);
Whenever you call the ToggleButton objects OnStateChange method, notify triggers the event:
tb.OnStateChange(true) ToggledState is true tb.OnStateChange(false) ToggledState is false
Removing Listeners
You can remove a listener object by calling delete on its handle. For example, if the class RespondToToggle above saved the listener handle as a property, you could delete the listener:
classdef RespondToToggle < handle properties ListenerHandle end
9-20
methods function obj = RespondToToggle(toggle_button_obj) hl = addlistener(toggle_button_obj,'ToggledState',@RespondToToggle.handleEvnt); obj.ListenerHandle = hl; end end ... end
With this code change, you can remove the listener from an instance of the RespondToToggle class. For example:
tb = ToggleButton; rtt = RespondToToggle(tb);
At this point, the object rtt is listening for the ToggleState event triggered by object tb. To remove the listener, call delete on the property containing the listener handle:
delete(rtt.ListenerHandle)
You do not need to explicitly delete a listener. MATLAB automatically deletes the listener when the objects lifecycle ends (e.g., when the rtt object is deleted). See Limiting Listener Scope Constructing event.listener Objects Directly on page 9-23 for related information.
Defining Event-Specific Data

Suppose that you want to pass to the listener callback the state of the toggle button as a result of the event. You can add more data to the default event data by subclassing the event.EventData class and adding a property to contain this information. You then can pass this object to the notify method.
classdef ToggleEventData < event.EventData properties NewState end methods
9-21
function data = ToggleEventData(newState) data.NewState = newState; end end end
The call to notify uses the ToggleEventData constructor to create the necessary argument.
notify(obj,'ToggledState',ToggleEventData(newState));
Ways to Create Listeners

When you call the notify method, the MATLAB runtime sends the event data to all registered listener callbacks. There are two ways to create a listener: Use the addlistener method, which binds the listener to the lifecycle of the object(s) that will generate the event. The listener object persists until the object it is attached to is destroyed. Use the event.listener class constructor. In this case, the listeners you create are not tied to the lifecycle of the object(s) being listened to. Instead the listener is active so long as the listener object remains in scope and is not deleted.
Attach Listener to Event Source Using addlistener

The following code defines a listener for the ToggleState event:
lh = addlistener(obj,'ToggleState',@CallbackFunction)
The arguments are: obj The object that is the source of the event ToggleState The event name passed as a string @CallbackFunction A function handle to the callback function The listener callback function must accept at least two arguments, which are automatically passed by the MATLAB runtime to the callback. The arguments are:
9-22
The source of the event (that is, obj in the call to addlistener) An event.EventData object, or a subclass of event.EventData , such as the ToggleEventData object described earlier Defining Event-Specific Data on page 9-21. The callback function must be defined to accept these two arguments:
function CallbackFunction(src,evnt) ... end
In cases where the event data (evnt) object is user defined, it must be constructed and passed as an argument to the notify method. For example, the following statement constructs a ToggleEventData object and passes it to notify as the third argument:
notify(obj,'ToggledState',ToggleEventData(newState));
Defining Listener Callback Functions on page 9-24 provides more information on callback syntax.
Limiting Listener Scope Constructing event.listener Objects Directly

You can also create listeners by calling the event.listener class constructor directly. When you call the constructor instead of using addlistener to create a listener, the listener exists only while the listener object you create is in scope (e.g., within the workspace of an executing function). It is not tied to the event-generating objects existence. The event.listener constructor requires the same arguments as used by addlistener the event-naming object, the event name, and a function handle to the callback:
lh = event.listener(obj,'ToggleState',@CallbackFunction)
If you want the listener to persist beyond the normal variable scope, you should use addlistener to create it.
9-23
Temporarily Deactivating Listeners

The addlistener method returns the listener object so that you can set its properties. For example, you can temporarily disable a listener by setting its Enabled property to false:
lh.Enabled = false;
To re-enable the listener, set Enabled to true. Enabling and Disabling the Listeners on page 9-51 provides an example.
Permanently Deleting Listeners

Calling delete on a listener object destroys it and permanently removes the listener:
delete(lh) % Listener object is removed and destroyed
Defining Listener Callback Functions

Callbacks are functions that execute when the listener receives notification of an event. Typically, you define a method in the class that creates the listener as the callback function. Pass a function handle that references the method to addlistener or the event.listener constructor when creating the listener.
function_handle provides more information on function handles.
All callback functions must accept at least two arguments: The handle of the object that is the source of the event An event.EventData object or an object that is derived from the event.EventData class (see Defining Event-Specific Data on page 9-21 for an example that extends this class).
Callback Syntax
For an function:
@functionName
For an ordinary method called with an object of the class:
9-24
@obj.methodName
For an static method:

@ClassName.methodName
Adding Arguments to a Callback Function

Ordinary class methods (i.e., not static methods) require a class object as an argument, so you need to add another argument to the callback function definition. If your listener callback is a method of the class of an object, obj, then your call to addlistener would use this syntax:
hlistener = addlistener(eventSourceObj,'MyEvent',@obj.listenMyEvent)
Another syntax uses an anonymous function. See the Anonymous Functions section for general information on anonymous functions For example, create a method to use as your callback function and reference this method as a function handle in a call to addlistener or the event.listener constructor:
hlistener = addlistener(eventSourceObj,'MyEvent',@(src,evnt)listenMyEvent(obj,src,evnt))
Then define the method in a method block as usual:

methods function listenMyEvent(obj,src,evnt) % obj - instance of this class % src - object generating event % evnt - the event data ... end end
Variables in the Expression provides information on variables used in anonymous functions.
9-25
Callback Execution
Listeners execute their callback function when notified that the event has occurred. Listeners are passive observers in the sense that errors in the execution of a listener callback does not prevent the execution of other listeners responding to the same event, or execution of the function that triggered the event. Callback function execution continues until the function completes. If an error occurs in a callback function, execution stops and control returns to the calling function. Then any remaining listener callback functions execute.
Listener Order of Execution

The order in which listeners callback functions execute after the firing of an event is undefined. However, all listener callbacks execute synchronously with the event firing.
Managing Callback Errors

If you want to control how your program responds to error, use a try/catch statement in your listener callback function to handle errors. See Respond to an Exception and the MException class.
9-26
Listen for Changes to Property Values

In this section... Creating Property Listeners on page 9-27 Property Event and Listener Classes on page 9-29 Aborting Set When Value Does Not Change on page 9-31
Creating Property Listeners

You can listen to the predeclared property events (named: PreSet, PostSet, PreGet, and PostGet) by creating a listener for those named events: Specify the SetObservable and/or GetObservable property attributes to add listeners for set or get events. Define a callback function Create a property listener by including the name of the property as well as the event in the call to addlistener (see Add a Listener to the Property on page 9-29.) Optionally subclass event.data to create a specialized event data object to pass to the callback function. Prevent execution of the callback if the new value is the same as the current value (see Aborting Set When Value Does Not Change on page 9-31).
Set Property Attributes to Enable Property Events

In the properties block, enable the SetObservable attribute:
properties (SetObservable) % Can define PreSet and PostSet property listeners % for properties defined in this block PropOne PropTwo ... end
9-27
Define a Callback Function for the Property Event

The listener executes the callback function when MATLAB triggers the property event. You must define the callback function to have two specific arguments, which are passed to the function automatically when called by the listener: Event source a meta.property object describing the object that is the source of the property event Event data a event.PropertyEvent object containing information about the event You can pass additional arguments if necessary. It is often simple to define this method as Static because these two arguments contain most necessary information in their properties. For example, suppose the handlePropEvents function is a static method of the class creating listeners for two properties of an object of another class:
methods (Static) function handlePropEvents(src,evnt) switch src.Name % switch on the property name case 'PropOne' % PropOne has triggered an event ... case 'PropTwo' % PropTwo has triggered an event ... end end end
Another possibility is to use the event.PropertyEvent objects EventName property in the switch statement to key off the event name (PreSet or PostSet in this case). Class Metadata on page 14-2 provides more information about the
meta.property class.
9-28
Add a Listener to the Property

The addlistener handle class method enables you to attach a listener to a property without storing the listener object as a persistent variable. For a property events, use the four-argument version of addlistener. If the call
addlistener(EventObject,'PropOne','PostSet',@ClassName.handlePropertyEvents);
The arguments are: EventObject handle of the object generating the event PropOne name of the property to which you want to listen PostSet name of the event for which you want to listen @ClassName.handlePropertyEvents function handle referencing a static method, which requires the use of the class name If your listener callback is an ordinary method and not a static method, the syntax is:
addlistener(EventObject,'PropOne','PostSet',@obj.handlePropertyEvents);
where obj is the handle of the object defining the callback method. If the listener callback is a function that is not a class method, you pass a function handle to that function. Suppose the callback function is a package function:
addlistener(EventObject,'PropOne','PostSet',@package.handlePropertyEvents);
See function_handle for more information on passing functions as arguments.
Property Event and Listener Classes

The following two classes show how to create PostSet property listeners for two properties PropOne and PropTwo.
9-29
Class Generating the Event

The PropEvent class enables property PreSet and PostSet event triggering by specifying the SetObservable property attribute. These properties also enable the AbortSet attribute, which prevents the triggering of the property events if the properties are set to a value that is the same as their current value (see Aborting Set When Value Does Not Change on page 9-31)
classdef PropEvent < handle % enable property events with the SetObservable attribute properties (SetObservable, AbortSet) PropOne PropTwo end methods function obj = PropEvent(p1,p2) if nargin > 0 obj.PropOne = p1; obj.PropTwo = p2; end end end end
Class Defining the Listeners

The PropListener class defines two listeners: Property PropOne PostSet event Property PropTwo PostSet event You could define listeners for other events or other properties using a similar approach and it is not necessary to use the same callback function for each listener. See the meta.property and event.PropertyEvent reference pages for more on the information contained in the arguments passed to the listener callback function.
classdef PropListener < handle methods function obj = PropListener(evtobj) % Pass the object generating the event to the constructor
9-30
% Add the listeners from the constructor if nargin > 0 addlistener(evtobj,'PropOne','PostSet',@PropListener.handlePropEvents); addlistener(evtobj,'PropTwo','PostSet',@PropListener.handlePropEvents); end end end methods (Static) function handlePropEvents(src,evnt) switch src.Name case 'PropOne' fprintf(1,'PropOne is %s\n',num2str(evnt.AffectedObject.PropOne)) case 'PropTwo' fprintf(1,'PropTwo is %s\n',num2str(evnt.AffectedObject.PropTwo)) end end end end
Aborting Set When Value Does Not Change

By default, MATLAB triggers the property PreSet and PostSet events, invokes the propertys set method (if defined), and sets the property value, even when the current value of the property is the same as the new value. You can prevent this behavior by setting the propertys AbortSet attribute to true. When AbortSet is true, MATLAB does not: Set the property value Trigger the PreSet and PostSet events Call the propertys set method, if one exists When AbortSet is true, MATLAB gets the current property value to compare it to the value you are assigning to the property. This causes the property get method (get.Property) to execute, if one exists. However, MATLAB does not catch errors resulting from the execution of this method and these errors are visible to the user.
9-31
When to Use AbortSet

Consider using AbortSet only when the cost of setting a property value is much greater than the cost of always comparing the current value of the property with the new value being assigned.
How AbortSet Works

The following example shows how the AbortSet attribute works. The AbortTheSet class defines a property, PropOne, that has listeners for the PreGet and PreSet events and enables the AbortSet attribute. The behavior of the post set/get events is equivalent so only the pre set/get events are used for simplicity: Note Save the AbortTheSet class in a file with the same name in a folder on your MATLAB path.
classdef AbortTheSet < handle properties (SetObservable, GetObservable, AbortSet) PropOne = 7 end methods function obj = AbortTheSet(val) obj.PropOne = val; addlistener(obj,'PropOne','PreGet',@obj.getPropEvt); addlistener(obj,'PropOne','PreSet',@obj.setPropEvt); end function propval = get.PropOne(obj) disp('get.PropOne called') propval = obj.PropOne; end function set.PropOne(obj,val) disp('set.PropOne called') obj.PropOne = val; end function getPropEvt(obj,src,evnt) disp ('Pre-get event triggered') end function setPropEvt(obj,src,evnt)
9-32
disp ('Pre-set event triggered') end function disp(obj) % Overload disp to avoid accessing property disp (class(obj)) end end end
The class specifies an initial value of 7 for the PropOne property. Therefore, if you create an object with the property value of 7, there is not need to trigger the PreSet event:
>> ats = AbortTheSet(7); get.PropOne called
If you specify a value other than 7, then MATLAB triggers the PreSet event:
>> ats = AbortTheSet(9); get.PropOne called set.PropOne called get.PropOne called
Similarly, if you set the PropOne property to the value 9, the AbortSet attribute prevents the property assignment and the triggering of the PreSet event. Notice also, that there is no PreGet event generated. Only the property get method is called:
>> ats.PropOne = 9; get.PropOne called
If you query the property value, the PreGet event is triggered:

>> a = ats.PropOne Pre-get event triggered get.PropOne called a = 9
If you set the PropOne property to a different value, MATLAB:
9-33
Calls the property get method to determine if the value is changing Triggers the PreSet event Calls the property set method to set the new value Calls the property get method again to determine if the result of calling the set method changed the value.
>> ats.PropOne = 11; get.PropOne called Pre-set event triggered set.PropOne called get.PropOne called
Because a property set method might modify the value that is actually assigned to a property, MATLAB must query the property value that would result from an assignment after a call the propertys set method. This results in multiple calls to a property get method, if one is defined for that property.
9-34
Update Graphs Using Events and Listeners

In this section... Example Overview on page 9-35 Access Fully Commented Example Code on page 9-36 Techniques Demonstrated in This Example on page 9-37 Summary of fcneval Class on page 9-37 Summary of fcnview Class on page 9-38 Methods Inherited from Handle Class on page 9-40 Using the fcneval and fcnview Classes on page 9-40 Implementing the UpdateGraph Event and Listener on page 9-43 The PostSet Event Listener on page 9-48 Enabling and Disabling the Listeners on page 9-51
Example Overview
This example defines two classes: fcneval The function evaluator class contains a MATLAB expression and evaluates this expression over a specified range fcnview The function viewer class contains a fcneval object and displays surface graphs of the evaluated expression using the data contained in fcneval. This class defines two events: A class-defined event that occurs when a new value is specified for the MATLAB function A property event that occurs when the property containing the limits is changed The following diagram shows the relationship between the two objects. The
fcnview object contains a fcneval object and creates graphs from the data it
9-35
contains. fcnview creates listeners to change the graphs if any of the data in the fcneval object change.
fcnview Properties fcneval object graph fcneval Properties FofXY Lm observable Data Events UpdateGraph
Listeners Lm property UpdateGraph
Access Fully Commented Example Code

You can display the code for this example in a popup window that contains detailed comments and links to related sections of the documentation by clicking these links:
fcneval class fcnview class createViews static method
You can open all files in your editor by clicking this link: Open in editor To use the classes, save the files in folders with the following names:
9-36
@fcneval/fcneval.m @fcnview/fcnview.m @fcnview/createViews.m The @-folders parent folder must be on the MATLAB path.
Techniques Demonstrated in This Example

Naming an event in the class definition Triggering an event by calling notify Enabling a property event via the SetObservable attribute Creating listeners for class-defined events and property PostSet events Defining listener callback functions that accept additional arguments Enabling and disabling listeners
Summary of fcneval Class

The fcneval class is designed to evaluate a MATLAB expression over a specified range of two variables. It is the source of the data that is graphed as a surface by instances of the fcnview class. It is the source of the events used in this example. Property
FofXY Lm
Value function handle
Purpose MATLAB expression (function of two variables).
two-element Limits over which function is evaluated in vector both variables. SetObservable attribute set to true to enable property event listeners. structure with x, y, and z matrices Data resulting from evaluating the function. Used for surface graph. Dependent attribute set to true, which means the get.Data method is called to determine property value when queried and no data is stored.
Data
9-37
Event
UpdateGraph
When Triggered
FofXY property set function (set.FofXY) calls the notify method when a new value is specified for the MATLAB expression on an object of this class.
Method
fcneval
Purpose Class constructor. Inputs are function handle and two-element vector specifying the limits over which to evaluate the function.
FofXY property set function. Called whenever property
set.FofXY set.Lm get.Data
value is set, including during object construction.

Lm property set function. Used to test for valid limits. Data property get function. This method calculates the values for the Data property whenever that data is queried (by class members or externally).
grid
A static method (Static attribute set to true) used in the calculation of the data.
Summary of fcnview Class

Instances of the fcnview class contain fcneval objects as the source of data for the four surface graphs created in a function view. fcnview creates the listeners and callback functions that respond to changes in the data contained in fcneval objects. Property
FcnObject HAxes
Value
fcneval object
Purpose This object contains the data that is used to create the function graphs. Each instance of a fcnview object stores the handle of the axes containing its subplot.
axes handle
9-38
Property
Value object for

UpdateGraph
Purpose Setting the event.listener objects Enabled property to true enables the listener; false disables listener. Setting the event.listener objects Enabled property to true enables the listener, false disables listener. Item on context menu used to enable listeners (used to handle checked behavior) Item on context menu used to disable listeners (used to manage checked behavior) Used by event callbacks to update surface data.
HLUpdateGraph event.listener
event
HLLm event.listener
object for Lm property event

HEnableCm
uimenu handle
HDisableCm
uimenu handle
HSurface
surface handle
Method
fcnview createLisn lims updateSurfaceData listenUpdateGraph listenLm delete createViews
Purpose Class constructor. Input is fcneval object. Calls addlistener to create listeners for UpdateGraph and Lm property PostSet listeners. Sets axes limits to current value of fcneval objects Lm property. Used by event handlers. Updates the surface data without creating a new object. Used by event handlers. Callback for UpdateGraph event. Callback for Lm property PostSet event Delete method for fcnview class. Static method that creates an instance of the fcnview class for each subplot, defines the context menus that enable/disable listeners, and creates the subplots
9-39
Methods Inherited from Handle Class

Both the fcneval and fcnview classes inherit methods from the handle class. The following table lists only those inherited methods used in this example. Handle Class Methods on page 5-12 provides a complete list of methods that are inherited when you subclass the handle class. Methods Inherited from Handle Class
addlistener notify
Purpose
Register a listener for a specific event and attach listener to event-defining object. Trigger an event and notify all registered listeners.
Using the fcneval and fcnview Classes

This sections explains how to use the classes. Create an instance of the fcneval class to contain the MATLAB expression of a function of two variables and the range over which you want to evaluate this function Use the fcnview class static function createViews to visualize the function Change the MATLAB expression or the limits contained by the fcneval object and all the fcnview objects respond to the events generated. You create a fcneval object by calling its constructor with two argumentsan anonymous function and a two-element, monotonically increasing vector. For example:
feobject = fcneval(@(x,y) x.*exp(-x.^2-y.^2),[-2 2]);
Use the createViews static method to create the graphs of the function. Note that you must use the class name to call a static function:
fcnview.createViews(feobject);
9-40
The createView method generates four views of the function contained in the fcneval object.
Each subplot defines a context menu that can enable and disable the listeners associated with that graph. For example, if you disable the listeners on subplot 221 (upper left) and change the MATLAB expression contained by the fcneval object, only the remaining three subplots update when the UpdateGraph event is triggered:
feobject.FofXY = @(x,y) x.*exp(-x.^.5-y.^.5);
9-41
Similarly, if you change the limits by assigning a value to the feobject.Lm property, the feobject triggers a PostSet property event and the listener callbacks update the graph.
feobject.Lm = [-8 3];
In this figure the listeners are re-enabled via the context menu for subplot 221. Because the listener callback for the property PostSet event also updates the surface data, all views are now synchronized
9-42
Implementing the UpdateGraph Event and Listener

The UpdateGraph event occurs when the MATLAB representation of the mathematical function contained in the fcneval object is changed. The fcnview objects that contain the surface graphs are listening for this event, so they can update the graphs to represent the new function.
Defining and Firing the UpdateGraph Event

The UpdateGraph event is a class-defined event. The fcneval class names the event and calls notify when the event occurs.
9-43
1. A property is assigned a new value. obj.FofXY = @(x,y)x^2+y^2 2. Setting the property runs a set access method, which, in turn, executes notify.
myfunceval Properties FofXY Methods set.FofXY Events UpdateGraph
function set.FofXY(obj,func) obj.FofXY = func; notify(obj,UpdateGraph); end
3. The notify method triggers an event, and a message is broadcast. 4. A listener awaiting the message executes its callback. UpdateGraph Listener Properties EventName = UpdateGraph FunctionHandle = @listenUpdateGraph
5. The callback function is executed.
myfuncview Methods listenUpdateGraph
The fcnview class defines a listener for this event. When fcneval triggers the event, the fcnview listener executes a callback function that performs the follow actions: Determines if the handle of the surface object stored by the fcnview object is still valid (that is, does the object still exist) Updates the surface XData, YData, and ZData by querying the fcneval objects Data property.
9-44
The fcneval class defines an event name in an event block:

events UpdateGraph end
Determining When to Trigger the Event

The fcneval class defines a property set method for the FofXY property. FofXY is the property that stores the MATLAB expression for the mathematical function. This expression must be a valid MATLAB expression for a function of two variables. The set.FofXY method: Determines the suitability of the expression If the expression is suitable:
Assigns the expression to the FofXY property Triggers the UpdateGraph event
If fcneval.isSuitable does not return an MException object, the set.FofXY method assigns the value to the property and triggers the UpdateGraph event.
function set.FofXY(obj,func) % Determine if function is suitable to create a surface me = fcneval.isSuitable(func); if ~isempty(me) throw(me) end % Assign property value obj.FofXY = func; % Trigger UpdateGraph event notify(obj,'UpdateGraph'); end
Determining Suitability of the Expression

The set.FofXY method calls a static method (fcneval.isSuitable) to determine the suitability of the specified expression. fcneval.isSuitable
9-45
returns an MException object if it determines that the expression is unsuitable. fcneval.isSuitable calls the MException constructor directly to create more useful error messages for the user.
set.FofXY issues the exception using the MException throw method. Issuing the exception terminates execution of set.FofXY and prevents the method from making an assignment to the property or triggering the UpdateGraph event.
Here is the fcneval.isSuitable method:

function isOk = isSuitable(funcH) v = [1 1;1 1]; % Can the expression except 2 numeric inputs try funcH(v,v); catch %#ok<CTCH> me = MException('DocExample:fcneval',... ['The function ',func2str(funcH),' Is not a suitable F(x,y)']); isOk = me; return end % Does the expression return non-scalar data if isscalar(funcH(v,v)); me = MException('DocExample:fcneval',... ['The function ',func2str(funcH),'' Returns a scalar when evaluated']); isOk = me; return end isOk = []; end
The fcneval.isSuitable method could provide additional test to ensure that the expression assigned to the FofXY property meets the criteria required by the class design.
Other Approaches
The class could have implemented a property set event for the FofXY property and would, therefore, not need to call notify (see Listen for Changes
9-46
to Property Values on page 9-27). Defining a class event provides more flexibility in this case because you can better control event triggering. For example, suppose you wanted to update the graph only if the new data is significantly different. If the new expression produced the same data within some tolerance, the set.FofXY method could not trigger the event and avoid updating the graph. However, the method could still set the property to the new value.
Defining the Listener and Callback for the UpdateGraph Event

The fcnview class creates a listener for the UpdateGraph event using the addlistener method:
obj.HLUpdateGraph = addlistener(obj.FcnObject,'UpdateGraph',... @(src,evnt)listenUpdateGraph(obj,src,evnt)); % Add obj to argument list
The fcnview object stores a handle to the event.listener object in its HLUpdateGraph property, which is used to enable/disable the listener by a context menu (see Enabling and Disabling the Listeners on page 9-51). The fcnview object (obj) is added to the two default arguments (src, evnt) passed to the listener callback. Keep in mind, the source of the event (src) is the fcneval object, but the fcnview object contains the handle of the surface object that is updated by the callback. The listenUpdateGraph function is defined as follows:
function listenUpdateGraph(obj,src,evnt) if ishandle(obj.HSurface) % If surface exists obj.updateSurfaceData % Update surface data end end
The updateSurfaceData function is a class method that updates the surface data when a different mathematical function is assigned to the fcneval object. Updating a graphics object data is generally more efficient than creating a new object using the new data:
function updateSurfaceData(obj)
9-47
% Get data from fcneval object and set surface data set(obj.HSurface,... 'XData',obj.FcnObject.Data.X,... 'YData',obj.FcnObject.Data.Y,... 'ZData',obj.FcnObject.Data.Matrix); end
The PostSet Event Listener

All properties support the predefined PostSet event (See Property-Set and Query Events on page 9-14 for more information on property events). This example uses the PostSet event for the fcneval Lm property. This property contains a two-element vector specifying the range over which the mathematical function is evaluated. Just after this property is changed (by a statement like obj.Lm = [-3 5];), the fcnview objects listening for this event update the graph to reflect the new data.
9-48
1. New limits are assigned. obj.Lm = [-3 5]; 2. The SetObservable attribute of Properties is set to True, so setting the property automatically triggers a PostSet event. Note that methods and events did not have to be declared in myfunceval. 3. A message is broadcast. 4. A listener awaiting the message executes its callback. PostSet
myfunceval Properties (SetObservable) Lm
Listener Properties EventName = PostSet FunctionHandle = @listenLm
5. The callback function is executed.
myfuncview Methods listenLm
Sequence During the Lm Property Assignment

The fcneval class defines a set function for the Lm property. When a value is assigned to this property during object construction or property reassignment, the following sequence occurs:
1 An attempt is made to assign argument value to Lm property. 2 The set.Lm method executes to check whether the value is in appropriate
range if yes, it makes assignment, if no, it generates an error.

3 If the value of Lm is set successfully, the MATLAB runtime triggers a
PostSet event.
9-49
4 All listeners execute their callbacks, but the order is nondeterministic.
The PostSet event does not occur until an actual assignment of the property occurs. The property set function provides an opportunity to deal with potential assignment errors before the PostSet event occurs.
Enabling the PostSet Property Event

To create a listener for the PostSet event, you must set the propertys SetObservable attribute to true:
properties (SetObservable = true) Lm = [-2*pi 2*pi]; % specifies default value end
The MATLAB runtime automatically triggers the event so it is not necessary to call notify. Specifying Property Attributes on page 6-7 provides a list of all property attributes.
Defining the Listener and Callback for the PostSet Event

The fcnview class creates a listener for the PostSet event using the addlistener method:
obj.HLLm = addlistener(obj.FcnObject,'Lm','PostSet',... @(src,evnt)listenLm(obj,src,evnt)); % Add obj to argument list
The fcnview object stores a handle to the event.listener object in its HLLm property, which is used to enable/disable the listener by a context menu (see Enabling and Disabling the Listeners on page 9-51). The fcnview object (obj) is added to the two default arguments (src, evnt) passed to the listener callback. Keep in mind, the source of the event (src) is the fcneval object, but the fcnview object contains the handle of the surface object that is updated by the callback.
9-50
The callback sets the axes limits and updates the surface data because changing the limits causes the mathematical function to be evaluated over a different range:
function listenLm(obj,src,evnt) if ishandle(obj.HAxes) % If there is an axes lims(obj); % Update its limits if ishandle(obj.HSurface) % If there is a surface obj.updateSurfaceData % Update its data end end end
Enabling and Disabling the Listeners

Each fcnview object stores the handle of the listener objects it creates so that the listeners can be enabled or disabled via a context menu after the graphs are created. All listeners are instances of the event.listener class, which defines a property called Enabled. By default, this property has a value of true, which enables the listener. If you set this property to false, the listener still exists, but is disabled. This example creates a context menu active on the axes of each graph that provides a way to change the value of the Enabled property.
Context Menu Callback

There are two callbacks used by the context menu corresponding to the two items on the menu: Listen Sets the Enabled property for both the UpdateGraph and PostSet listeners to true and adds a check mark next to the Listen menu item. Dont Listen Sets the Enabled property for both the UpdateGraph and PostSet listeners to false and adds a check mark next to the Dont Listen menu item. Both callbacks include the fcnview object as an argument (in addition to the required source and event data arguments) to provide access to the handle of the listener objects. The enableLisn function is called when the user selects Listen from the context menu.
9-51
function enableLisn(obj,src,evnt) obj.HLUpdateGraph.Enabled = true; % Enable listener obj.HLLm.Enabled = true; % Enable listener set(obj.HEnableCm,'Checked','on') % Check Listen set(obj.HDisableCm,'Checked','off') % Uncheck Don't Listen end
The disableLisn function is called when the user selects Dont Listen from the context menu.
function disableLisn(obj,src,evnt) obj.HLUpdateGraph.Enabled = false; % Disable listener obj.HLLm.Enabled = false; % Disable listener set(obj.HEnableCm,'Checked','off') % Unheck Listen set(obj.HDisableCm,'Checked','on') % Check Don't Listen end
9-52
10
Building on Other Classes
Hierarchies of Classes Concepts on page 10-2 Creating Subclasses Syntax and Techniques on page 10-7 Modifying Superclass Methods and Properties on page 10-14 Subclassing Multiple Classes on page 10-18 Controlling Allowed Subclasses on page 10-20 Controlling Access to Class Members on page 10-24 Supporting Both Handle and Value Subclasses on page 10-34 Subclassing MATLAB Built-In Types on page 10-43 Determining the Class of an Array on page 10-73 Defining Abstract Classes on page 10-77 Defining Interfaces on page 10-82
10
Hierarchies of Classes Concepts

In this section... Classification on page 10-2 Developing the Abstraction on page 10-3 Designing Class Hierarchies on page 10-4 Super and Subclass Behavior on page 10-4 Implementation and Interface Inheritance on page 10-5
Classification
Organizing classes into hierarchies facilitates the reuse of code and the reuse of solutions to design problems that have already been solved. You can think of class hierarchies as sets supersets (referred to as superclasses or base classes), and subsets (referred to as subclasses or derived classes). For example, the following picture shows how you could represent an employee database with classes.
10-2
Sales People and Engineers are subsets of Employees
Base class Employees Properties Name Address Department
Employees
Sales People
Engineers Derived classes Test Engineers SalesPerson (is an Employees) Properties Commission Region Engineer (is an Employees) Properties Products Team
TestEngineer (is an Engineer) Properties TestStage

At the root of the hierarchy is the Employees class. It contains data and operations that apply to the set of all employees. Contained in the set of employees are subsets whose members, while still employees, are also members of sets that more specifically define the type of employee. Subclasses like TestEngineer are examples of these subsets.
Developing the Abstraction

Classes are representations of real world concepts or things. When designing a class, form an abstraction of what the class represents. Consider an abstraction of an employee and what are the essential aspects of employees
10-3
10
for the intended use of the class. Name, address, and department can be what all employees have in common. When designing classes, your abstraction should contain only those elements that are necessary. For example, the employee hair color and shoe size certainly characterize the employee, but are probably not relevant to the design of this employee class. Their sales region is relevant only to some employee so this characteristic belongs in a subclass.
Designing Class Hierarchies

As you design a system of classes, put common data and functionality in a superclass, which you then use to derive subclasses. The subclasses inherit the data and functionality of the superclass and define only aspects that are unique to their particular purposes. This approach provides advantages: Avoid duplicating code that is common to all classes. Add or change subclasses at any time without modifying the superclass or affecting other subclasses. If the superclass changes (for example, all employees are assigned a number), then subclass automatically get these changes.
Super and Subclass Behavior

Subclass objects behave like objects of the superclass because they are specializations of the superclass. This fact facilitates the development of related classes that behave similarly, but are implemented differently.
A Subclass Object Is A Superclass Object

You can usually describe the relationship between an object of a subclass and an object of its superclass with a statement like: The subclass is a superclass . For example: An Engineer is an Employee. This relationship implies that objects belonging to a subclass have the same properties, methods, and events as the superclass, as well as any new features defined by the subclass. Test this relationship with the isa function.
10-4
Treat Subclass Objects Like Superclass Objects

You can pass a subclass object to a superclass method, but you can access only those properties that the superclass defines. This behavior enables you to modify the subclasses without affecting the superclass. Two points about super and subclass behavior to keep in mind are: Methods defined in the superclass can operate on subclass objects. Methods defined in the subclass cannot operate on superclass objects. Therefore, you can treat an Engineer object like any other Employees object, but an Employee object cannot pass for an Engineer object.
Limitations to Object Substitution

MATLAB determines the class of an object based on its most specific class. Therefore, an Engineer object is of class Engineer, while it is also an Employees object, as using the isa function reveals. Generally, MATLAB does not allow you to create arrays containing a mix of superclass and subclass objects because an array can be of only one class. If you attempt to concatenate objects of different classes, MATLAB looks for a converter method defined by the less dominant class (usually, the left-most object in the expression is the dominant class). See Concatenating Objects of Different Classes on page 8-13 for more information. See matlab.mixin.Heterogeneous for information on defining heterogeneous class hierarchies. See Converting Objects to Another Class on page 15-7 for information on defining converter methods.
Implementation and Interface Inheritance

MATLAB classes support both the inheritance of implemented methods from a superclass and the inheritance of interfaces defined by abstract methods in the superclass.
10-5
10
Implementation inheritance enables code reuse by subclasses. For example, an employee class can have a submitStatus method that all employee subclasses can use. Subclasses can extend an inherited method to provide specialized functionality, while reusing the common aspects. See Modifying Superclass Methods and Properties on page 10-14 for more information on this process. Interface inheritance is useful in cases where you want a group of classes to provide a common interface, but these classes create specialized implementations of methods and properties that define the interface. Create an interface using an abstract class as the superclass. This class defines the methods and properties that you must implement in the subclasses, but does not provide an implementation. The subclasses must provide their own implementation of the abstract members of the superclass. To create an interface, define methods and properties as abstract using their Abstract attribute. See Defining Abstract Classes on page 10-77 for more information and an example.
10-6
Creating Subclasses Syntax and Techniques

In this section... Defining a Subclass on page 10-7 Initializing Superclasses from Subclasses on page 10-7 Constructor Arguments and Object Initialization on page 10-10 Call Only Direct Superclass from Constructor on page 10-10 Sequence of Constructor Calls in a Class Hierarchy on page 10-12 Using a Subclass to Create an Alias for an Existing Class on page 10-12
Defining a Subclass
To define a class that is a subclass of another class, add the superclass to the classdef line after a < character:
classdef classname < superclassname
When inheriting from multiple classes, use the & character to indicate the combination of the superclasses:
classdef classname < super1 & super2
See Class Member Compatibility on page 10-18 for more information on deriving from multiple superclasses.
Class Attributes
Subclasses do not inherit superclass attributes.
Initializing Superclasses from Subclasses

Use the following syntax to initialize the object for each superclass within the subclass constructor.
obj@baseclass1(args); ...
10-7
10
obj@baseclassN(args);
Where obj is the output of the constructor, baseclass is the name of a superclass, and args are any arguments required by the respective superclass constructor. For example, the following segment of a class definition shows a class called stock that is a subclass of a class called asset.
classdef stock < asset methods function s = stock(asset_args,...) if nargin == 0 ... end % Call asset constructor s = s@asset(asset_args); ... end end end
Constructing Subclasses on page 7-20 provides more information on creating subclass constructor methods.
Referencing Superclasses Contained in Packages

If you are deriving a class from a superclass that is contained in a package and you want to initialize the object for the superclass, include the package name. For example:
classdef stock < financial.asset methods function s = stock(asset_args,...) if nargin == 0 ... end % Call asset constructor s = [email protected](asset_args); ... end
10-8
end end
Initializing Objects When Using Multiple Superclasses

If you are deriving a class from multiple superclasses, initialize the subclass object with calls to each superclass constructor:
classdef stock < financial.asset & trust.member methods function s = stock(asset_args,member_args,...) if nargin == 0 ... end % Call asset and member class constructors s = [email protected](asset_args) s = [email protected](member_args) ... end end end
Explicitly calling each superclass constructor enables you to: Pass arguments to superclass constructors Control the order in which MATLAB calls the superclass constructors If you do not explicitly call the superclass constructors from the subclass constructor, MATLAB implicitly calls these constructors with no arguments. In this case, the superclass constructors must support no argument syntax. See No Input Argument Constructor Requirement on page 7-19 for more information. In the case of multiple superclasses, MATLAB does not guarantee any specific calling sequence. If the order in which MATLAB calls the superclass constructors is important, you must explicitly call the superclass constructors from the subclass constructor.
10-9
10
Constructor Arguments and Object Initialization

You cannot conditionalize calls to the superclass initialization of the object. Locate calls to superclass constructors outside any conditional code blocks. However, always ensure that your class constructor supports the zero arguments syntax. You can satisfy the need for a zero-argument syntax by assigning appropriate values to input argument variables before constructing the object: For example, the stock class constructor supports the no argument case with the if statement, but initializes the object for the superclass outside of the if code block.
classdef stock < financial.asset properties SharePrice end methods function s = stock(name,pps) % Support no input argument case if nargin == 0 name = ''; pps = 0; end % Call superclass constructor s = [email protected](name) % Assign property value s.SharePrice = pps; end end end
See No Input Argument Constructor Requirement on page 7-19.
Call Only Direct Superclass from Constructor

You cannot call an indirect superclass constructor from a subclass constructor. For example, suppose class B derives from class A and class C derives from class B. The constructor for class C cannot call the constructor for class A to initialize properties. Class B must make the call to initialize class A properties.
10-10
The following implementations of classes A, B, and C show how to design this relationship in each class. Class A defines properties x and y, but assigns a value only to x:
classdef A properties x y end methods function obj = A(x) obj.x = x; end end end
Class B inherits properties x and y from class A. The class B constructor calls the class A constructor to initialize x and then assigns a value to y.
classdef B < A methods function obj = B(x,y) obj = obj@A(x); obj.y = y; end end end
Class C accepts values for the properties x and y, and passes these values to the class B constructor, which in turn calls the class A constructor:
classdef C < B methods function obj = C(x,y) obj = obj@B(x,y); end end end
10-11
10
Sequence of Constructor Calls in a Class Hierarchy

MATLAB always calls the most specific subclass constructor first to enable you to call superclass constructors explicitly. Suppose you have a hierarchy of class in which ClassC derives from ClassB, which derives from ClassA:
ClassA
ClassB
ClassC
MATLAB always calls the most specific class constructor (ClassC in this case) first. This approach enables you to process input arguments and perform any necessary setup before calling the superclass constructors. If you do not make an explicit call to a superclass constructor from the subclass constructor, MATLAB makes the implicit call before accessing the object. The order is always from most specific to least specific and all the superclass constructors must finish executing before the subclass can access the object. You can change the order in which class constructors are called by calling superclass constructors explicitly from the subclass constructor.
Using a Subclass to Create an Alias for an Existing Class

You can refer to a class using a different name by creating an alias for that class. This technique is like the C++ typedef concept. To create an alias, create an empty subclass:
classdef newclassname < oldclassname end
10-12
The old class constructor must be callable with zero input arguments. If not, see Old Class Constructor Requires Arguments on page 10-13. This technique is useful when reloading objects that you saved using the old class name. However, the class of the object reflects the new name. For example,
class(obj)
returns the new class name.
Old Class Constructor Requires Arguments

If the old class constructor requires arguments, add a constructor to the new class:
classdef NewClass < OldClass methods function obj = NewClass(x,y) obj = obj@OldClass(x,y); end end
10-13
10
Modifying Superclass Methods and Properties

In this section... Modifying Superclass Methods on page 10-14 Modifying Superclass Properties on page 10-16 Private Local Property Takes Precedence in Method on page 10-16
Modifying Superclass Methods

An important concept in class design is that a subclass object is also an object of its superclass. Therefore, you can pass a subclass object to a superclass method and have the method execute properly. At the same time, you can apply special processing to the unique aspects of the subclass. Some useful techniques include: Calling a superclass method from within a subclass method Redefining in the subclass protected methods called from within a public superclass method Defining the same named methods in both super and subclass, but using different implementations
Extending Superclass Methods

Subclass methods can call superclass methods of the same name. This fact enables you to extend a superclass method in a subclass without completely redefining the superclass method. For example, suppose that both superclass and subclass defines a method called foo. The method names are the same so the subclass method can call the superclass method. However, the subclass method can also perform other steps before and after the call to the superclass method. It can operate on the specialized parts to the subclass that are not part of the superclass. For example, this subclass defines a foo method, which calls the superclass
foo method classdef sub < super methods
10-14
function foo(obj)
preprocessing steps
foo@super(obj); % Call superclass foo method

postprocessing steps
end end end
See Invoking Superclass Methods in Subclass Methods on page 7-14 for more on this syntax.
Completing Superclass Methods

A superclass method can define a process that executes in a series of steps using a protected method for each step (Access attribute set to protected). Subclasses can then create their own versions of the protected methods that implement the individual steps in the process. Implement this technique as shown here:
classdef super methods function foo(obj) step1(obj) step2(obj) step3(obj) end end methods (Access = protected) function step1(obj)
superclass version
end ... end end
The subclass does not reimplement the foo method, it reimplements only the methods that carry out the series of steps (step1(obj), step2(obj), step3(obj)). That is, the subclass can specialize the actions taken by each step, but does not control the order of the steps in the process. When you pass
10-15
10
a subclass object to the superclass foo method, MATLAB calls the subclass step methods because of the dispatching rules.
classdef sub < super ... methods (Access = protected) function step1(obj)
subclass version
end ... end end
Redefining Superclass Methods

You can completely redefine a superclass method. In this case, both the superclass and the subclass would define the same named method.
Modifying Superclass Properties

There are two separate conditions under which you can redefine superclass properties: The value of the superclass property Abstract attribute is true The values of the superclass property SetAccess and GetAccess attributes are private In the first case, the superclass is just requesting that you define a concrete version of this property to ensure a consistent interface. In the second case, only the superclass can access the private property, so the subclass is free to reimplement it in any way.
Private Local Property Takes Precedence in Method

When a subclass property has the same name as a superclass private property, and a method of the superclass references the property name, MATLAB always accesses the property defined by the calling methods class. For example, given the following classes, Sub and Super:
classdef Super
10-16
properties (Access = private) Prop = 2; end methods function p = superMethod(obj) p = obj.Prop; end end end classdef Sub < Super properties Prop = 1; end end
If you create an instance of the subclass and use it to call the superclass method, MATLAB access the private property of the methods class:
subObj = Sub subObj = Sub with properties: Prop: 1 subObj.superMethod ans = 2
10-17
10
Subclassing Multiple Classes

In this section... Class Member Compatibility on page 10-18 Using Multiple Inheritance on page 10-19
Class Member Compatibility

When you create a subclass derived from multiple classes, the subclass inherits the properties, methods, and events defined by all specified superclasses. If more than one superclass defines a property, method, or event having the same name, there must be an unambiguous resolution to the multiple definitions. You cannot derive a subclass from any two or more classes that define incompatible class members. There are various situations where you can resolve name and definition conflicts, as described in the following sections.
Property Conflicts
If two or more superclasses define a property with the same name, then at least one of the following must be true: All, or all but one of the properties must have their SetAccess and GetAccess attributes set to private The properties have the same definition in all superclasses (for example, when all superclasses inherited the property from a common base class)
Method Conflicts
If two or more superclasses define methods with the same name, then at least one of the following must be true: The methods Access attribute is private so only the defining superclass can access the method. The method has the same definition in all subclasses. This situation can occur when all superclasses inherit the method from a common base class and none of the superclasses override the inherited definition.
10-18
Subclassing Multiple Classes
The subclass redefines the method to disambiguate the multiple definitions across all superclasses. This means that the superclass methods must not have their Sealed attribute set to true. Only one superclass defines the method as Sealed, in which case, the subclass adopts the sealed method definition. The superclasses define the methods as Abstract and rely on the subclass to define the method.
Event Conflicts
If two or more superclasses define events with the same name, then at least one of the following must be true: The events ListenAccess and NotifyAccess attributes must be private. The event has the same definition in all superclasses (for example, when all superclasses inherited the event from a common base class)
Using Multiple Inheritance

Resolving the potential conflicts involved when defining a subclass from multiple classes often reduces the value of this approach. For example, problems can arise when you enhance superclasses in future versions and introduce new conflicts. Reduce potential problems by implementing only one unrestricted superclass. In all other superclasses, all methods are abstract and must be defined by a subclass or inherited from the unrestricted superclass. In general, when using multiple inheritance, ensure that all superclasses remain free of conflicts in definition. See Defining a Subclass on page 10-7 for the syntax used to derive a subclass from multiple superclasses. See Supporting Both Handle and Value Subclasses on page 10-34 for techniques that provide greater flexibility when using multiple superclasses.
10-19
10
Controlling Allowed Subclasses

In this section... Basic Knowledge on page 10-20 Why Control Allowed Subclasses on page 10-20 Specify Allowed Subclasses on page 10-21 Define a Sealed Hierarchy of Classes on page 10-22
Basic Knowledge
The material presented in this section builds on an understanding of the following information: Class Metadata on page 14-2 Specifying Attributes on page 3-23
Why Control Allowed Subclasses

A class definition can specify a list of classes that it allows to subclass the class. Classes not in the list cannot subclass the class. Use the AllowedSubclasses class attribute to specify the allowed subclasses. The AllowedSubclasses attribute provides a design point between Sealed classes, which do not allow subclassing, and the default behavior, which does not restrict subclassing. By controlling the allowed subclasses, you can create a sealed hierarchy of classes. That is, a system of classes that enables a specific set of classes to derive from specific base classes, but that does not allow unrestricted subclassing. See Define a Sealed Hierarchy of Classes on page 10-22 for more about this technique.
10-20
Specify Allowed Subclasses

Note Specify attribute values explicitly, not as values returned from functions or other MATLAB expressions. Specify a list of one or more allowed subclasses in the classdef statement by assigning meta.class objects to the AllowedSubclasses attribute. Create the meta.class object referencing a specific class using the ? operator and the class name:
classdef (AllowedSubclasses = ?ClassName) MySuperClass ... end
Use a cell array of meta.class objects to define more than one allowed subclass:
classdef (AllowedSubclasses = {?ClassName1,?ClassName2,...?ClassNameN}) MySuperClass ... end
Always use the fully qualified class name when referencing the class name:
classdef (AllowedSubclasses = ?Package.SubPackage.ClassName1) MySuperClass ... end
Assigning an empty cell array to the AllowedSubclasses attribute is effectively the same as defining a Sealed class.
classdef (AllowedSubclasses = {}) MySuperClass ... end
10-21
10
Note Use only the ? operator and the class name to generate meta.class objects. Values assigned to the AllowedSubclasses attribute cannot contain any other MATLAB expressions, including functions that return either meta.class objects or cell arrays of meta.class objects.
Effect of Defining a Class as an Allowed Subclass

Including a class in the list of AllowedSubclasses does not define that class as a subclass or require you to define the class as a subclass. It just allows the referenced class to be defined as a subclass. Declaring a class as an allowed subclass does not affect whether this class can itself be subclassed. A class definition can contain assignments to the AllowedSubclasses attribute that reference classes that are not currently defined or available on the MATLAB path. However, any referenced subclass that MATLAB cannot find when loading the class is effectively removed from the list without causing an error or warning. Note If MATLAB does not find any of the classes in the allowed classes list, the class is effectively Sealed. This is equivalent to AllowedSubclasses = {}.
Define a Sealed Hierarchy of Classes

The AllowedSubclasses attribute enables you to define a sealed class hierarchy by sealing the allowed subclasses:
classdef (AllowedSubclasses = {?SubClass1,?SubClass2}) SuperClass ... end
Define the allowed subclasses as Sealed:

classdef (Sealed) SubClass1 ...
10-22
end classdef (Sealed) SubClass2 ... end
Sealed class hierarchies enable you to use the level of abstraction that your design requires while maintaining a closed systems of classes. See Supporting Both Handle and Value Subclasses on page 10-34 for related information.
10-23
10
Controlling Access to Class Members

In this section... Basic Knowledge on page 10-24 Applications for Access Control Lists on page 10-25 Specify Access to Class Members on page 10-26 Properties with Access Lists on page 10-29 Methods with Access Lists on page 10-29 Abstract Methods with Access Lists on page 10-33
Basic Knowledge
The material presented in this section builds on an understanding of the following information: Related Topics Class Metadata on page 14-2 Specifying Attributes on page 3-23 Terminology and Concepts Class members Properties, methods, and events defined by a class Defining class The class defining the class member for which access is being specified Get access Permission to read the value of a property; controlled by the property GetAccess attribute Set access Permission to assign a value to a property; controlled by the property SetAccess attribute Method access Determines what other methods and functions can call the class method; controlled by the method Access attribute Listen access Permission to define listeners; controlled by the event ListenAccess attribute
10-24
Notify access Permission to trigger events; controlled by the event NotifyAccess attribute Possible Values for Access to Class Members The following class member attributes can contain a list of classes: Properties Access, GetAccess, and SetAccess. See Property Attributes on page 6-8 for a list of all property attributes. Methods Access. See Method Attributes on page 7-5 for a list of all method attributes. Events ListenAccess and NotifyAccess. See Event Attributes on page 9-16 for a list of all event attributes. These attributes accept the following possible values: public Unrestricted access protected Access by defining class and its subclasses private Access by defining class only Access list A list of one or more classes. Only the defining class and the classes in the list have access to the class members to which the attribute applies. If you specify a list of classes, MATLAB does not allow access by any other class (that is, access is private, except for the listed classes).
Applications for Access Control Lists

Access control lists enable you to control access to specific class properties, methods, and events, by specifying a list of classes to which you want to grant access to these class members. This technique provides greater flexibility and control in the design of a system of classes. For example, use access control lists when you want to define parts of your class system in separate classes, but do not want to allow access to class members from outside the class system.
10-25
10
Specify Access to Class Members

Specify the classes that are allowed to access a particular class member in the member access attribute statement. For example:
methods (Access = {?ClassName1,?ClassName2,...})
Use the class meta.class object to refer to classes in the access list. To specify more than one class, use a cell array of meta.class objects. Use the package name when referring to classes that are in packages. Note You must specify the meta.class objects explicitly (created with the ? operator), not as values returned by functions or other MATLAB expressions.
Property Access The following class declares access lists for the property GetAccess and Access attributes:
classdef PropertyAccess properties (GetAccess = {?ClassA, ?ClassB}, SetAccess = private) Prop1 end properties (Access = ?ClassC) Prop2 end end
The class PropertyAccess specifies the following property access: Gives the classes ClassA and ClassB get access to the Prop1 property. Gives all subclasses of ClassA and ClassB get access to the Prop1 property. Does not give get access to Prop1 from subclasses of PropertyAccess. Defines private set access for the Prop1 property. Gives set and get access to Prop2 for ClassC and its subclasses.
10-26
Method Access The following class declares an access list for the method Access attribute:
classdef MethodAccess methods (Access = {?ClassA, ?ClassB, ?MethodAccess}) function listMethod(obj) ... end end end
The MethodAccess class specifies the following method access: Access to listMethod from an instance of MethodAccess by methods of the classes ClassA and ClassB. Access to listMethod from an instance of MethodAccess by methods of subclasses of MethodAccess, because of the inclusion of MethodAccess in the access list. Subclasses of ClassA and ClassB are allowed to define a method named listMethod, and MethodAccess is allowed to redefine listMethod. However, if MethodAccess was not in the access list, its subclasses could not redefine listMethod. Event Access The following class declares an access list for the event ListenAccess attribute:
classdef EventAccess events (NotifyAccess = private, ListenAccess = {?ClassA, ?ClassB}) Event1 Event2 end end
The class EventAccess specifies the following event access: Limits notify access for Event1 and Event2 to EventAccess; only methods of the EventAccess can trigger these events.
10-27
10
Gives listen access for Event1 and Event2 to methods of ClassA and ClassB. Methods of EventAccess, ClassA, and ClassB can define listeners for these events. Subclasses of MyClass cannot define listeners for these events. See Methods with Access Lists on page 10-29.
How MATLAB Interprets Attribute Values

Granting access to a list of classes restricts access to only:
The defining class The classes in the list Subclasses of the classes in the list
Including the defining class in the access list gives all subclasses of the defining class access. MATLAB resolves references to classes in the access list only when the class is loaded. If MATLAB cannot find a class that is included in the access list, that class is effectively removed from the list. An empty access list (that is, an empty cell array) is equivalent to private access.
Specifying Metaclass Objects

Use only the ? operator and the class name to generate the meta.class objects. Values assigned to the attributes cannot contain any other MATLAB expressions, including functions that return allowed attribute values: meta.class objects Cell arrays of meta.class objects The values public, protected, or private You must specify these values explicitly, as shown in the example code in this section.
10-28
Properties with Access Lists

These sample classes show the behavior of a property that grants read access (GetAccess) to a class. The GrantAccess class gives GetAccess to the NeedAccess class for the Prop1 property:
classdef GrantAccess properties (GetAccess = ?NeedAccess) Prop1 = 7; end end
The NeedAccess class defines a method that uses the value of the GrantAccess Prop1 value. The dispObj is defined as a Static method, however, it could be an ordinary method.
classdef NeedAccess methods (Static) function dispObj(GrantAccessObj) % Display the value of Prop1 disp(['Prop1 is: ',num2str(GrantAccessObj.Prop1)]) end end end
Get access to Prop1 is private so MATLAB returns an error:

>> a = GrantAccess; >> a.Prop1 Getting the 'Prop1' property of the 'GrantAccess' class is not allowed.
However, MATLAB allows access to Prop1 by the NeedAccess class:

>> NeedAccess.dispObj(a) Prop1 is: 7
Methods with Access Lists

Classes granted access to a method can: Call the method using an instance of the defining class. Define their own method with the same name (if not a subclass).
10-29
10
Override the method in a subclass only if the superclass defining the method includes itself or the subclass in the access list. These sample classes show the behavior of methods called from methods of other classes that are in the access list. The class AcListSuper gives the AcListNonSub class access to itsm1 method:
classdef AcListSuper methods (Access = {?AcListNonSub}) function obj = m1(obj) disp ('Method m1 called') end end end
Because AcListNonSub is in the access list of m1, its methods can call m1 using an instance of AcListSuper:
classdef AcListNonSub methods function obj = nonSub1(obj,AcListSuper_Obj) % Call m1 on AcListSuper class AcListSuper_Obj.m1; end function obj = m1(obj) % Define a method named m1 disp(['Method m1 defined by ',class(obj)]) end end end
Create objects of both classes:

>> a = AcListSuper; >> b = AcListNonSub;
Call the AcListSuper m1 method using an AcListNonSub method:

>> b.nonSub1(a); Method m1 called
10-30
Call the AcListNonSub m1 method:

>> b.m1; Method m1 defined by AcListNonSub
Subclasses Without Access

Including the defining class in the access list for a method grants access to all subclasses derived from that class. When you derive from a class that has a method with an access list and that list does not include the defining class in the access list: Subclass methods cannot call the superclass method because it is effectively private. Subclasses cannot override the superclass method. Subclass methods can call the superclass method indirectly using an instance of a class that is in the access list. Nonsubclass methods of classes in the superclass method access list can call the superclass method using an instance of a subclass that is not in the superclass method access list. For example, AcListSub is a subclass of AcListSuper. The AcListSuper class defines an access list for method m1. However, this list does not include AcListSuper, which would implicitly include all subclasses of AcListSuper in the access list:
classdef AcListSub < AcListSuper methods function obj = sub1(obj,AcListSuper_Obj) % Access m1 via superclass object (NOT ALLOWED) AcListSuper_Obj.m1; end function obj = sub2(obj,AcListNonSub_Obj,AcListSuper_obj) % Access m1 via object that is in access list (is allowed) AcListNonSub_Obj.nonSub1(AcListSuper_Obj); end end end
10-31
10
Attempting to call the superclass m1 method results in an error because subclasses are not in the access list for the method:
>> a = AcListSuper; >> b = AcListNonSub; >> c = AcListSub; >> c.sub1(a); Error using AcListSuper/m1 Cannot access method 'm1' in class 'AcListSuper'. Error in AcListSub/sub1 (line 4) AcListSuper_Obj.m1;
The AcListSub sub2 method can call a method of a class that is on the access list for m1, and that method (nonSub1) does have access to the superclass m1 method:
>> c.sub2(b,a); Method m1 called
When subclasses are not included in the access list for a method, those subclasses cannot define a method with the same name. This behavior is not the same as cases in which the methods Access is explicitly declared as private. For example, adding the following method to the AcListSub class definition produces an error when you attempt to instantiate the class.
methods (Access = {?AcListNonSub}) function obj = m1(obj) disp('AcListSub m1 method') end end
If you attempt to instantiate the class, MATLAB returns an error:

>> c = AcListSub; Error using AcListSub Class 'AcListSub' is not allowed to override the method 'm1' because neither it nor its superclasses have been granted access to the method by class 'AcListSuper'.
10-32
The AcListNonSub class, which is in the m1 method access list, can define a method that calls the m1 method using an instance of the AcListSub class. While AcListSub is not in the access list for method m1, it is a subclass of AcListSuper. For example, add the following method to the AcListNonSub class:
methods function obj = nonSub2(obj,AcListSub_Obj) disp('Call m1 via subclass object:') AcListSub_Obj.m1; end end
Calling the nonSub2 method results in execution of the superclass m1 method:

>> b = AcListNonSub; >> c = AcListSub; >> b.nonSub2(c); Call m1 via subclass object: Method m1 called
This is consistent with the behavior of any subclass object, which can be substituted for an instance of its superclass.
Abstract Methods with Access Lists

A class containing a method declared as Abstract is an abstract class. It is the responsibility of subclasses to implement the abstract method using the function signature declared in the class definition. When an abstract method has an access list, only the classes in the access list can implement the method. A subclass that is not in the access list cannot implement the abstract method so that subclass is itself abstract.
10-33
10
Supporting Both Handle and Value Subclasses

In this section... Basic Knowledge on page 10-34 Handle Compatibility Rules on page 10-34 Defining Handle-Compatible Classes on page 10-35 Subclassing Handle-Compatible Classes on page 10-38 Methods for Handle Compatible Classes on page 10-39 Handle-Compatible Classes and Heterogeneous Arrays on page 10-41
Basic Knowledge
The material presented in this section builds on knowledge of the following information. Creating Subclasses Syntax and Techniques on page 10-7 Subclassing Multiple Classes on page 10-18 Comparing Handle and Value Classes on page 5-2
Key Concepts
Handle-compatible class is a class that you can combine with handle classes when defining a set of superclasses. All handle classes are handle-compatible. All superclasses of handle-compatible classes must also be handle compatible.
HandleCompatible the class attribute that defines nonhandle classes as
handle compatible.
Handle Compatibility Rules

Handle-compatible classes (that is, classes whose HandleCompatible attribute is set to true) follow these rules:
10-34
All superclasses of a handle-compatible class must also be handle compatible If a class explicitly sets its HandleCompatibility attribute to false, then none of the classs superclasses can be handle classes. If a class does not explicitly set its HandleCompatible attribute and, if any superclass is a handle, then all superclasses must be handle compatible. The HandleCompatible attribute is not inherited. A class that does not explicitly set its HandleCompatible attribute to true is: A handle class if any of its superclasses are handle classes A value class if none of the superclasses are handle classes
Defining Handle-Compatible Classes

A class is handle compatible if: It is a handle class Its HandleCompatible attribute is set to true The HandleCompatible class attribute identifies classes that you can combine with handle classes when specifying a set of superclasses. Handle compatibility provides greater flexibility when defining abstract superclasses, such as mixin and interface classes, in cases where the superclass is designed to support both handle and value subclasses. Handle compatibility removes the need to define both a handle version and a nonhandle version of a class.
A Handle Compatible Class

For example, consider a class named Utility that defines functionality that is useful to both handle and value subclasses. In this example, the Utility class defines a method to reset property values to the default values defined in the respective class definition:
classdef (HandleCompatible) Utility methods
10-35
10
function obj = resetDefaults(obj) % Reset properties to default and return object mc = metaclass(obj); % Get meta.class object mp = mc.PropertyList; % Get meta.property objects for k=1:length(mp) % For each property, if there is a default defined, % set the property to that value if mp(k).HasDefault && ~strcmp(mp(k).SetAccess,'private') obj.(mp(k).Name) = mp(k).DefaultValue; end end end end end
The Utility class is handle compatible. Therefore, you can use it in the derivation of classes that are either handle classes or value classes. See Getting Information About Classes and Objects for information on using meta-data classes.
Return Modified Objects

The resetDefaults method defined by the Utility class returns the object it modifies, which is necessary when you call resetDefaults with a nonhandle object. It is important to implement methods that work with both handle and value objects in a handle compatible superclass. See Modifying Objects on page 3-51 for more information on modifying handle and value objects. Consider the behavior of a value class that subclasses the Utility class. The PropertyDefaults class defines three properties, all of which have default values:
classdef PropertyDefaults < Utility properties p1 = datestr(rem(now,1)); % Current time p2 = 'red'; % Character string p3 = pi/2; % Result of division operation end end
10-36
Create a PropertyDefaults object. MATLAB evaluates the expressions assigned as default property values when the class is first loaded, and uses these same default values whenever you create an instance of this class in the current MATLAB session.
pd = PropertyDefaults with properties: p1: ' 4:54 PM' p2: 'red' p3: 1.5708
Assign new values that are different from the default values:
pd.p1 = datestr(rem(now,1)); pd.p2 = 'green'; pd.p3 = pi/4;
All pd object property values now contain values that are different from the default values originally defined by the class:
pd = PropertyDefaults with properties: : p1: ' 5:36 PM' p2: 'green' p3: 0.7854
Call the resetDefaults method, which is inherited from the Utility class. Because the PropertyDefaults class is not a handle class, you must return the modified object for reassignment in the calling functions workspace.
pd = pd.resetDefaults pd = PropertyDefaults with properties: p1: ' 4:54 PM' p2: 'red' p3: 1.5708
10-37
10
If the PropertyDefaults class was a handle class, then you would not need to save the object returned by the resetDefaults method. However, to design a handle compatible class like Utility, you need to ensure that all methods work with both kinds of classes.
Subclassing Handle-Compatible Classes

According to the rules described in Handle Compatibility Rules on page 10-34, when you combine a handle superclass with a handle-compatible superclass, the result is a handle subclass, which is handle compatible. However, subclassing a handle-compatible class does not necessarily result in the subclass being handle compatible. Consider the following two cases, which demonstrate two possible results.
Combine Nonhandle Utility Class with Handle Classes

Suppose you define a class that subclasses a handle class, as well as the handle compatible Utility class discussed in A Handle Compatible Class on page 10-35. The HPropertyDefaults class has these characteristics: It is a handle class (it derives from handle). All of its superclasses are handle compatible (handle classes are handle compatible by definition).
classdef HPropertyDefaults < handle & Utility properties GraphPrim = line; Width = 1.5; Color = 'black'; end end
The HPropertyDefaults class is handle compatible:

hpd = HPropertyDefaults; mc = metaclass(hpd); mc.HandleCompatible ans =
10-38
Nonhandle Subclasses of a Handle-Compatible Class

If you subclass a value class that is not handle compatible in combination with a handle compatible class, the subclass is a nonhandle compatible value class. The ValueSub class: Is a value class (It does not derive from handle.) One of its superclasses is handle compatible (the Utility class).
classdef ValueSub < MException & Utility % ValueSub class is-a value class that is not % itself a handle-compatibile class methods function obj = ValueSub(str1,str2) obj = obj@MException(str1,str2); end end end
The ValueSub class is a nonhandle-compatible value class because the MException class does not define the HandleCompatible attribute as true:
hv = ValueSub('MATLAB:narginchk:notEnoughInputs',... 'Not enough input arguments.'); mc = metaclass(hv); mc.HandleCompatible ans = 0
Methods for Handle Compatible Classes

Objects passed to methods of handle compatible classes can be either handle or value objects. There are two different behaviors to consider when implementing methods for a class that operate on both handles and values:
10-39
10
If an input object is a handle object, then the method can alter the handle object and these changes are visible to all workspaces that have the same handle. If an input object is a value object, then changes to the object made inside the method affect only the value inside the method workspace. Handle compatible methods generally do not alter input objects because the effect of such changes are not the same for handle and nonhandle objects. See Modifying Objects on page 3-51 for information about modifying handle and value objects.
Identifying Handle Objects

Use the isa function to determine if an object is a handle object:
isa(obj,'handle')
Modifying Value Objects in Methods

If a method operates on both handle and value objects, the method must return the modified object. For example, the TimeStamp property returns the object it modifies:
classdef (HandleCompatible) Util % Utility class that adds a time stamp properties TimeStamp end methods function obj = setTime(obj) % Return object after modification obj.TimeStamp = now; end end end
10-40
Handle-Compatible Classes and Heterogeneous Arrays

A heterogeneous array contains objects of different classes. Members of a heterogeneous array have a common superclass, but might belong to different subclasses. See the matlab.mixin.Heterogeneous class for more information on heterogeneous arrays. The matlab.mixin.Heterogeneous class is a handle-compatible class.
Methods Must Be Sealed

You can invoke only those methods that are sealed by the common superclass on heterogeneous arrays (Sealed attribute set to true). Sealed methods prevent subclasses from overriding those methods and guarantee that methods called on heterogeneous arrays have the same definition for the entire array. Subclasses cannot override sealed methods. In situations requiring subclasses to specialize methods defined by a utility class, you can employ the design pattern referred to as the template method.
Using the Template Technique

Suppose you need to implement a handle compatible class that is intended to work with heterogeneous arrays. The following approach enables you to seal public methods, while providing a way for each subclass to specialize how the method works on each subclass instance: In the handle compatible class:
Define a sealed method that accepts a heterogeneous array as input. Define a protected, abstract method that each subclass must implement. Within the sealed method, call the overridden method for each array element.
Each subclass in the heterogeneous hierarchy implements a concrete version of the abstract method, which provides specialized behavior required by the particular subclass. The Printable class shows how to implement a template method approach:
10-41
10
classdef (HandleCompatible) Printable methods(Sealed) function print(aryIn) % Print elements of a potentially % heterogeneous array n = numel(aryIn); for k=1:n % Call subclass concrete implementation printElement(aryIn(k)); end end end methods(Access=protected, Abstract) % Define protected, abstract method % Each subclass implements a concrete version printElement(objIn) end end
10-42
Subclassing MATLAB Built-In Types

In this section... MATLAB Built-In Types on page 10-43 Why Subclass Built-In Types on page 10-44 Which Functions Work With Subclasses of Built-In Types on page 10-44 Built-In Types You Cannot Subclass on page 10-44 Examples of Classes that Subclass of Built-In Types on page 10-45 Behavior of Built-In Functions with Subclass Objects on page 10-45 A Class to Manage uint8 Data on page 10-51 Subclasses of Built-In Types with Properties on page 10-59 Understanding size and numel on page 10-65 A Class to Represent Hardware on page 10-70
MATLAB Built-In Types

Built-in types represent fundamental kinds of data such as numeric arrays, logical arrays, and character arrays. Other built-in types contain data belonging to these fundamental types and other classes. For example, cell and struct arrays can contain instances of any class. Built-in types define methods that perform operations on objects of these classes. For example, you can perform operations on numeric arrays, such as, sorting, rounding values, and element-wise and matrix multiplication. You can create an object of class double using an assignment statement, indexing expressions, or using converter functions. See Fundamental MATLAB Classes for more information on MATLAB built-in classes. Note It is an error to define a class that has the same name as a built-in class.
10-43
10
Why Subclass Built-In Types

Subclass a built-in type to extend the operations that you can perform on a particular class of data. For example, when you want to: Define unique operations to perform on class data. Be able to use methods of the built-in class and other built-in functions directly with objects of the subclass. For example, you do not need to reimplement all the mathematical operators if you derived from a class (for example, double) that defines these operators. See Built-In Types You Cannot Subclass on page 10-44 for a list of which MATLAB built-in classes you can subclass.
Which Functions Work With Subclasses of Built-In Types

Consider a class that defines enumerations. It can derive from an integer class and inherit methods that enable you to compare and sort values. For example, integer classes like int32 support all the relational methods (eq, ge, gt, le, lt, ne). To see a list of functions that the subclass has inherited as methods, use the methods function:
methods('SubclassName')
Generally, you can use an object of the subclass with any of the inherited methods and any functions coded in MATLAB that normally accept input arguments of the same class as the superclass. See Behavior of Built-In Functions with Subclass Objects on page 10-45 for information on other required methods.
Built-In Types You Cannot Subclass

You cannot subclass the following built-in MATLAB classes: char cell
10-44
struct function_handle
Examples of Classes that Subclass of Built-In Types

A Class to Manage uint8 Data on page 10-51 Subclasses of Built-In Types with Properties on page 10-59 A Class to Represent Hardware on page 10-70
Behavior of Built-In Functions with Subclass Objects

When you define a subclass of a built-in class, the subclass inherits all built-in class methods. In addition, MATLAB provide a number of built-in functions as subclass methods. However, built-in functions that work on built-in classes behave differently with subclasses, depending on which function you are using and whether your subclass defines properties.
Behavior Categories
When you call an inherited method on a subclass of a built-in class, the result of that call depends on the nature of the operation performed by the method. The behaviors of these methods fit into several categories. Operations on data values return objects of the superclass. For example, if you subclass double and perform addition on two subclass objects, MATLAB adds the numeric values and returns a value of class double. Operations on the orientation or structure of the data return objects of the subclass. Methods that perform these kinds of operations include, reshape, permute, transpose, and so on. Converting a subclass object to a built-in class returns an object of the specified class. Functions such as uint32, double, char, and so on, work with subclass objects the same as they work with built-in objects. Comparing objects or testing for inclusion in a specific set returns logical or built-in objects, depending on the function. Functions such as isequal, ischar, isobject, and so on work with subclass objects the same as they work with superclass objects.
10-45
10
Indexing expressions return objects of the subclass. If the subclass defines properties, then default indexing no longer works and the subclass must define its own indexing methods. See Subclasses That Define Properties on page 10-46 for more information. Concatenation returns an object of the subclass. If the subclass defines properties, then default concatenation no longer works and the subclass must define its own concatenation methods. See Subclasses That Define Properties on page 10-46 for more information. To list the built-in functions that work with a subclass of a built-in class, use the methods function.
Subclasses That Define Properties

When a subclass of a built-in class defines properties, MATLAB no longer provides support for indexing and concatenation operations. MATLAB cannot use the built-in functions normally called for these operations because subclass properties can contain any data. The subclass must define what indexing and concatenation mean for a class with properties. If your subclass needs indexing and concatenation functionality, then the subclass must implement the appropriate methods. The sections that follow list the methods you must implement in the subclass to support indexing and concatenation. Also, the section Subclasses of Built-In Types with Properties on page 10-59 provides an example of these methods. Methods for Concatenation. To support concatenation, the subclass must implement the following methods: horzcat Implement horizontal concatenation of objects vertcat Implement vertical concatenation of objects cat Implement concatenation of object arrays along specified dimension Concatenation Functions on page 10-50 Methods for Indexing. To support indexing operations, the subclass must implement these methods:
10-46
subsasgn Implement dot notation and indexed assignments subsref Implement dot notation and indexed references subsindex Implement object as index value Indexing Methods on page 10-50
More information on Built-In Methods

The following sections describe how different categories of methods behave with subclasses: Extending the Operations of a Built-In Class on page 10-47 Built-In Methods That Operate on Data Values on page 10-49 Built-In Methods That Operate on Data Organization on page 10-49
Extending the Operations of a Built-In Class

The MATLAB built-in class double defines a wide range of methods to perform arithmetic operations, indexing, matrix operation, and so on. Therefore, subclassing double enables you to add specific features without implementing many of the methods that a numeric class requires to function effectively in the MATLAB language. The following class definition subclasses the built-in class double.
classdef DocSimpleDouble < double methods function obj = DocSimpleDouble(data) if nargin == 0 data = 0; end obj = obj@double(data); % initialize the base class portion end end end
You can create an instance of the class DocSimpleDouble and call any methods of the double class.
10-47
10
sc = DocSimpleDouble(1:10) sc = 1x10 DocSimpleDouble: double data: 1 2 3 4
10
Calling a method inherited from class double that operates on the data, like sum, returns a double and, therefore, uses the display method of class double:
sum(sc) ans = 55
You can index sc like an array of doubles. The returned value is the class of the subclass, not double:
a = sc(2:4) a = 1x3 DocSimpleDouble: double data: 2 3 4
Indexed assignment also works:

sc(1:5) = 5:-1:1 sc = 1x10 DocSimpleDouble: double data: 5 4 3 2
10
Calling a method that modifies the order of the data elements operates on the data, but returns an object of the subclass:
sc = DocSimpleDouble(1:10); sc(1:5) = 5:-1:1; a = sort(sc) a = 1x10 DocSimpleDouble: double data: 1 2 3 4
10
10-48
Extending the Subclass. You can extend the DocSimpleDouble with specialized methods to provide custom behavior. For example, see A Class to Manage uint8 Data on page 10-51.
Built-In Methods That Operate on Data Values

Most built-in functions used with built-in classes are actually methods of the built-in class. For example, the double and single classes both have a sin method. All of these built-in class methods work with subclasses of the built-in class. When you call a built-in method on a subclass object, MATLAB uses the superclass part of the subclass object as inputs to the method, and the value returned is same class as the built-in class. For example:
sc = DocSimpleDouble(1:10); a = sin(sc); class(a) ans = double
Built-In Methods That Operate on Data Organization

This group of built-in methods reorders or reshapes the input argument array. These methods operate on the superclass part of the subclass object, but return an object of the same type as the subclass. Methods in this group include: reshape permute sort transpose ctranspose
10-49
10
Indexing Methods
Built-in classes use specially implemented versions of the subsref, subsasgn, and subsindex methods to implement indexing (subscripted reference and assignment). When you index a subclass object, only the built-in data is referenced (not the properties defined by your subclass). For example, indexing element 2 in the DocSimpleDouble subclass object returns the second element in the vector:
sc = DocSimpleDouble(1:10); a = sc(2) a = DocSimpleDouble double data: 2
The value returned from an indexing operation is an object of the subclass. You cannot make subscripted references if your subclass defines properties unless your subclass overrides the default subsref method. Assigning a new value to the second element in the DocSimpleDouble object operates only on the superclass data:
sc(2) = 12 sc = 1x10 DocSimpleDouble: double data: 1 12 3 4
10
The subsref method also implements dot notation for methods. See Subclasses of Built-In Types with Properties on page 10-59 for an example of a subsref method.
Concatenation Functions
Built-in classes use the functions horzcat, vertcat, and cat to implement concatenation. When you use these functions with subclass objects of the same type, MATLAB concatenates the superclass data to form a new object. For example, you can concatenate objects of the DocSimpleDouble class:
sc1 = DocSimpleDouble(1:10); sc2 = DocSimpleDouble(11:20);
10-50
[sc1,sc2] ans = 1x20 DocSimpleDouble: double data: Columns 1 through 13 1 14 [sc1; sc2] ans = 2x10 DocSimpleDouble: double data: 1 11 2 12 3 13 4 14 5 15 6 16 7 17 8 18 9 19 10 20 2 15 3 16 4 17 5 18 6 19 7 20 8 9 10 11 12 13 Columns 14 through 20
Concatenate two objects along a third dimension:

c = cat(3,sc1,sc2) c = cat(3,sc1,sc2) c = DocSimpleDouble double data: (:,:,1) = 1 2 3 (:,:,2) = 11 12 13
4 14
5 15
6 16
7 17
8 18
9 19
10 20
If the subclass of built-in class defines properties, you cannot concatenate objects of the subclass. Such an operation does not make sense because there is no way to know how to combine properties of different objects. However, your subclass can define custom horzcat and vertcat methods to support concatenation in whatever way makes sense for your subclass. See Concatenating DocExtendDouble Objects on page 10-64 for an example.
A Class to Manage uint8 Data

This example shows a class derived from the built-in uint8 class. This class simplifies the process of maintaining a collection of intensity image data defined by uint8 values. The basic operations of the class include:
10-51
10
Capability to convert various classes of image data to uint8 to reduce object data storage. A method to display the intensity images contained in the subclass objects. Ability to use all the methods that you can use on uint8 data (for example, size, indexing (reference and assignment), reshape, bitshift, cat, fft, arithmetic operators, and so on). The class data are matrices of intensity image data stored in the superclass part of the subclass object. This approach requires no properties. The DocUint8 class stores the image data, which converts the data, if necessary:
classdef DocUint8 < uint8 methods function obj = DocUint8(data) % Support no argument case if nargin == 0 data = uint8(0); % If image data is not uint8, convert to uint8 elseif ~strcmp('uint8',class(data)) switch class(data) case 'uint16' t = double(data)/65535; data = uint8(round(t*255)); case 'double' data = uint8(round(data*255)); otherwise error('Not a supported image class') end end % assign data to superclass part of object obj = obj@uint8(data); end % Get uint8 data and setup call to imagesc function h = showImage(obj) data = uint8(obj); figure; colormap(gray(256)) h = imagesc(data,[0 255]);
10-52
axis image brighten(.2) end end end
Using the DocUint8 Class

The DocUint8 class contains its own conversion code and provides a method to display all images stored as DocUint8 objects in a consistent way. For example:
cir = imread('circuit.tif'); img1 = DocUint8(cir); img1.showImage;
10-53
10
50
100
150
200
250
50
100
150
200
250
Because DocUint8 subclasses uint8, you can use any of its methods. For example,
size(img1) ans = 280 272
returns the size of the image data.
Indexing Operations
Inherited methods perform indexing operations, but return objects of the same class as the subclass.
10-54
Therefore, you can index into the image data and call a subclass method:
showImage(img1(100:200,1:160));
Subscripted reference operations (controlled by the inherited subsref method) return a DocUint8 object.
10 20 30 40 50 60 70 80 90 100 20 40 60 80 100 120 140 160
You can assign values to indexed elements:

img1(100:120,140:160) = 255; img1.showImage;
10-55
10
Subscripted assignment operations (controlled by the inherited subsasgn method) return a DocUint8 object.
50
100
150
200
250
50
100
150
200
250
Concatenation Operations
Concatenation operations work on DocUint8 objects because this class inherits the uint8 horzcat and vertcat methods, which return a DocUint8 object:
showImage([img1 img1]);
10-56
50 100 150 200 250 50 100 150 200 250 300 350 400 450 500
Data Operations
Methods that operate on data values, such as arithmetic operators, always return an object of the built-in type (not of the subclass type). For example, multiplying DocUint8 objects returns a uint8 object:
showImage(img1.*.8); ??? Undefined function or method 'showImage' for input arguments of type 'uint8'.
If you must be able to perform operations of this type, implement a subclass method to override the inherited method. The times method implements
10-57
10
array (element-by-element) multiplication. See Implementing Operators for Your Class on page 15-31 for a list of operator method names. For example:
function o = times(obj,val) u8 = uint8(obj).*val; % Call uint8 times method o = DocUint8(u8); end
Keep in mind that when you override a uint8 method, MATLAB calls the subclass method and no longer dispatches to the base class method. Therefore, explicitly call the uint8 times method or an infinite recursion can occur. Make the explicit call in this statement of the DocUint8 times method:
u8 = uint8(obj).*val;
After adding the times method to DocUint8, you can use the showImage method in expressions like:
showImage(img1.*1.8);
10-58
50
100
150
200
250
50
100
150
200
250
Subclasses of Built-In Types with Properties

When your subclass defines properties, indexing and concatenation do not work by default. There is really no way for the default subsref, horzcat, and vertcat methods to work with unknown property types and values. The following example subclasses the double class and defines a single property intended to contain a descriptive character string.
Methods Implemented
The following methods modify the behavior of the DocExtendDouble class:
10-59
10
DocExtendDouble The constructor supports a no argument syntax that initializes properties to empty values. subsref Enables subscripted reference to the superclass part (double) of the subclass, dot notation reference to the DataString property, and dot notation reference the built-in data via the string Data (the double data property is hidden). horzcat Defines horizontal concatenation of DocExtendDouble objects as the concatenation of the superclass part using the double class horzcat method and forms a cell array of the string properties. vertcat The vertical concatenation equivalent of hortzcat (both are required). char A DocExtendDouble to char converter used by horzcat and vertcat. disp DocExtendDouble implements a disp method to provide a custom display for the object.
Property Added
The DocExtendDouble class defines the DataString property to contain text that describes the data contained in instances of the DocExtendDouble class. Keep in mind that the superclass part (double) of the class contains the data.
Subclass with Properties

The DocExtendDouble class extends double and implements methods to support subscripted reference and concatenation.
classdef DocExtendDouble < double properties DataString end methods function obj = DocExtendDouble(data,str) % Support calling with zero arguments but do not return empty object if nargin == 0 data = 0;
10-60
str = ''; elseif nargin == 1 str = ''; end obj = obj@double(data); obj.DataString = str; end function sref = subsref(obj,s) % Implements dot notation for DataString and Data % as well as indexed reference switch s(1).type case '.' switch s(1).subs case 'DataString' sref = obj.DataString; case 'Data' sref = double(obj); if length(s)>1 && strcmp(s(2).type, '()') sref = subsref(sref,s(2:end)); end end case '()' sf = double(obj); if ~isempty(s(1).subs) sf = subsref(sf,s(1:end)); else error('Not a supported subscripted reference') end sref = DocExtendDouble(sf,obj.DataString); end end function newobj = horzcat(varargin) % Horizontal concatenation - cellfun calls double % on all object to get superclass part. cellfun call local char % to get DataString and the creates new object that combines % doubles in vector and chars in cell array and creates new object d1 = cellfun(@double,varargin,'UniformOutput',false ); data = horzcat(d1{:});
10-61
10
str = horzcat(cellfun(@char,varargin,'UniformOutput',false)); newobj = DocExtendDouble(data,str); end function newobj = vertcat(varargin) % Need both horzcat and vertcat d1 = cellfun(@double,varargin,'UniformOutput',false ); data = vertcat(d1{:}); str = vertcat(cellfun(@char,varargin,'UniformOutput',false)); newobj = DocExtendDouble(data,str); end function str = char(obj) % Used for cat functions to return DataString str = obj.DataString; end function disp(obj) % Change the default display disp(obj.DataString) disp(double(obj)) end end end
Create an instance of DocExtendDouble and notice that the display is different from the default:
ed = DocExtendDouble(1:10,'One to ten') ed = One to ten 1 2 3 4 5 6
10
The sum function continues to operate on the superclass part of the object:
sum(ed) ans = 55
Subscripted reference works because the class implements a subsref method:
10-62
ed(10:-1:1) ans = One to ten 10 9
However, subscripted assignment does not work because the class does not define a subsasgn method:
ed(1:5) = 5:-1:1 Error using DocExtendDouble/subsasgn Cannot use '(' or '{' to index into an object of class 'DocExtendDouble' be 'DocExtendDouble' defines properties and subclasses 'double'. Click here for more information.
The sort function works on the superclass part of the object:

sort(ed) ans = One to ten 1 2
10
Indexed Reference of a DocExtendDouble Object

Subscripted reference (performed by subsref) requires the subclass to implement its own subsref method.
ed = DocExtendDouble(1:10,'One to ten'); a = ed(2) a = One to ten 2 whos Name Size Bytes Class a 1x1 132 DocExtendDouble ed 1x10 204 DocExtendDouble
Attributes
You can access the property data:

c = ed.DataString
10-63
10
c = One to ten whos Name c ed
Size 1x10 1x10
Bytes 20 204
Class char DocExtendDouble
Concatenating DocExtendDouble Objects

Given the following two objects:
ed1 = DocExtendDouble([1:10],'One to ten'); ed2 = DocExtendDouble([10:-1:1],'Ten to one');
You can concatenate these objects along the horizontal dimension:

hcat = [ed1,ed2] hcat = 'One to ten' 1 7 whos Name ed1 ed2 hcat Size 1x10 1x10 1x20 Bytes 204 204 528 Class DocExtendDouble DocExtendDouble DocExtendDouble 2 6 3 5 'Ten to one' 4 4 5 3 6 2 7 1 8 9 10 10 9 8 Columns 1 through 13 Columns 14 through 20
Vertical concatenation works in a similar way:

vcat = [ed1;ed2] vcat = 'One to ten' 1 2 3 10 9 8
'Ten to one' 4 5 7 6
6 5
7 4
8 3
9 2
10 1
Both horzcat and vertcat return a new object of the same class as the subclass.
10-64
Understanding size and numel

The size function returns the dimensions of an array. The numel function returns the number of elements in an array. The default size and numel functions behave consistently with user-defined classes (see Classes Not Derived from Built-In Classes on page 10-67). Other MATLAB functions use size and numel to perform their operations and you usually do not need to overload them. When used with subclasses of built-in classes, the size and numel functions behave the same as in the superclasses. Consider the built-in class double:
d = 1:10; size(d) ans = 1 numel(d) ans = 10 dsubref = d(7:end); whos dsub Name dsubref 10
Size 1x4
Bytes 32
Class double
Attributes
The double class defines these behaviors, including parentheses indexing.
10-65
10
Subclass Inherited Behavior

Classes behave like their superclasses, unless the subclass explicitly overrides any given behavior. For example, DocSimpleDouble subclasses double, but defines no properties:
classdef DocSimpleDouble < double methods function obj = DocSimpleDouble(data) if nargin == 0 data = 0; end obj = obj@double(data); end end end
Create an object and assign to the superclass part of the object the values 1:10:
sd = DocSimpleDouble(1:10);
The size function returns the size of the superclass part:

size(sd) ans = 1 10
The numel function returns the number of elements in the superclass part:
numel(sd) ans = 10
Object arrays return the size of the built-in arrays also:

size([sd;sd])
10-66
ans = 2 10
numel([sd;sd]) ans = 20
The DocSimpleDouble class inherits the indexing behavior of the double class:
sdsubref = sd(7:end); whos sdsubref Name Size sdsubref 1x4
Bytes 88
Class DocSimpleDouble
Attributes
Classes Not Derived from Built-In Classes

Consider a simple value class. It does not inherit the array-like behaviors of the double class. For example:
classdef VerySimpleClass properties Value end end
Create an instance of this class and assign a ten-element vector to the Value property:
vs = VerySimpleClass; vs.Value = 1:10; size(vs) ans = 1 numel(vs) 1
10-67
10
ans = 1 size([vs;vs]) ans = 2 1
numel([vs;vs]) ans = 2 vs is a scalar object, as opposed to an array of VerySimpleClass objects. The Value property is an array of doubles: size(vs.Value) ans = 1 10
Apply indexing expressions to the object property:

vssubref = vs.Value(7:end); whos vssubref Name Size vssubref 1x4
Bytes 32
Class double
Attributes
vs.Value is an array of class double: class(vs.Value) ans = double
10-68
Creating an array of VerySimpleClass objects

vsArray(1:10) = VerySimpleClass;
MATLAB does not apply scalar expansion to object array property value assignment. Use the deal function for this purpose:
[vsArray.Value] = deal(1:10);
Indexing rules for object arrays are equivalent to those of struct arrays:
v1 = vsArray(1).Value; >> whos v1 Name Size v1 1x10
Bytes 80
Class double
Attributes
vsArray(1).Value(6) ans = 6
Changing the Behavior of size

Subclasses of built-in numeric classes inherit a size method, which operates on the superclass part of the subclass object (this method is hidden). If you want size to behave in another way, you can override it by defining your own size method in your subclass. Keep in mind that other MATLAB functions use the values returned by size. If you change the way size behaves, ensure that the values returned make sense for the intended use of your class.
Avoid Overloading numel

It is important to understand the significance of numel with respect to indexing. MATLAB calls numel to determine the number of elements returned by an indexed expression like:
A(index1,index2,...,indexn)
10-69
10
Both subsref and subsasgn use numel: subsref numel computes the number of expected outputs (nargout) returned subsref subsasgn numel computes the number of expected inputs (nargin) that MATLAB assigns as a result of a call to subsasgn Subclasses of built-in classes always return scalar objects as a result of subscripted reference and always use scalar objects for subscripted assignment. The numel function returns the correct value for these operations and there is, therefore, no reason to overload numel. If you define a class in which nargout for subsref or nargin for subsasgn is different from the value returned by the default numel, then overload numel for that class to ensure that it returns the correct values.
A Class to Represent Hardware

This example shows the implementation of a class to represent an optical multiplex card. These cards typically have a number of input ports, which this class represents by the port data rates and names. There is also an output port. The output rate of a multiplex card is the sum of the input port data rates. The DocMuxCard class defines the output rate as a Dependent property, and then defines a get access method for this property. The get.OutPutRate method calculates the actual output rate whenever the OutPutRate property is queried. See Property Get Methods on page 6-18 for more information on this technique.
Why Derive from int32

The DocMuxCard class derives from the int32 class because 32bit integers represent the input port data rates. The DocMuxCard class inherits the methods of the int32 class, which simplifies the implementation of this subclass.
10-70
Class Definition
Here is the definition of the DocMuxCard class. Notice that the input port rates initialize the int32 portion of class.
classdef DocMuxCard < int32 properties InPutNames % cell array of strings OutPutName % a string end properties (Dependent = true) OutPutRate end methods function obj = DocMuxCard(inptnames, inptrates, outpname) obj = obj@int32(inptrates); % initial the int32 class portion obj.InPutNames = inptnames; obj.OutPutName = outpname; end function x = get.OutPutRate(obj) x = sum(obj); % calculate the value of the property end function x = subsref(card, s) if strcmp(s(1).type,'.') base = subsref@int32(card, s(1)); if isscalar(s) x = base; else x = subsref(base, s(2:end)); end else x = subsref(int32(card), s); end end end end
Using the Class with Methods of int32

The constructor takes three arguments:
10-71
10
inptnames Cell array of input port names inptrates Vector of input port rates outpname Name for the output port
>> omx = DocMuxCard({'inp1','inp2','inp3','inp4'},[3 12 12 48],'outp') omx = 1x4 DocMuxCard array with properties: InPutNames: {'inp1' OutPutName: 'outp' OutPutRate: 75 int32 data: 3 12 12 48 'inp2' 'inp3' 'inp4'}
You can treat an DocMuxCard object like an int32. For example, this statement accesses the int32 data in the object to determine the names of the input ports that have a rate of 12:
>> omx.InPutNames(omx==12) ans = 'inp2' 'inp3'
Indexing the DocMuxCard object accesses the int32 vector of input port rates:
>> omx(1:2) ans = 3 12
The OutPutRate property get access method uses sum to sum the output port rates:
>> omx.OutPutRate ans = 75
10-72
Determining the Class of an Array

In this section... Querying the Class Name on page 10-73 Testing for Class on page 10-73 Testing for Specific Types on page 10-74 Testing for Most Derived Class on page 10-75
Querying the Class Name

Use the class function to determine the class of an array:
a = [2,5,7,11]; class(a) ans = double str = 'Character string'; class(str) ans = char
Testing for Class

The isa function enables you to test for a specific class or a category of numeric class (numeric, float, integer):
a = [2,5,7,11]; isa(a,'double') ans = 1
Floating-point values (single and double precision values):

isa(a,'float') ans = 1
Numeric values (floating-point and integer values):
10-73
10
isa(a,'numeric') ans = 1
isa Returns True for Subclasses

isa returns true for classes derived from the specified class. For example, the SubInt class derives from the built-in type int16: classdef SubInt < int16 methods function obj = SubInt(data) if nargin == 0 data = 0; end obj = obj@int16(data); end end end
By definition, an instance of the SubInt class is also an instance of the int16 class:
aInt = SubInt; isa(aInt,'int16') ans = 1
Using the integer category also returns true:

isa(aInt,'integer') ans = 1
Testing for Specific Types

The class function returns the name of the most derived class of an object:
class(aInt) ans = SubInt
10-74
Use the strcmp function with the class function to check for a specific class of an object:
a = int16(7); strcmp(class(a),'int16') ans = 1
Because the class function returns the class name as a character string, the inheritance of objects does not affect the result of the string comparison performed by strcmp:
aInt = SubInt; strcmp(class(aInt),'int16') ans = 0
Testing for Most Derived Class

If you define functions that require inputs that are: MATLAB built-in types Not subclasses of MATLAB built-in types Use the following techniques to exclude subclasses of built-in types from the input arguments. Define a cell array that contain the names of built-in types accepted by your function. Call class and strcmp to test for specific types in a MATLAB control statement. Test an input argument:
if strcmp(class(inputArg),'single') % Call function else inputArg = single(inputArg); end
10-75
10
Testing for a Category of Types

Suppose you create a MEX-function, myMexFcn, that requires two numeric inputs that must be of type double or single:
outArray = myMexFcn(a,b)
Define a cell array floatTypes contains the strings double and single:
floatTypes = {'double','single'}; % Test for proper types if any(strcmp(class(a),floatTypes)) && ... any(strcmp(class(b),floatTypes)) outArray = myMexFcn(a,b); else % Try to convert inputs to avoid error ... end
Another Test for Built-In Types

You can use isobject to separate built-in types from subclasses of built-in types. The isobject function returns false for instances of built-in types:
% Create a int16 array a = int16([2,5,7,11]); isobject(a) ans = 0
Determine if an array is one of the built-in integer types:

if isa(a,'integer') && ~isobject(a) % a is a built-in integer type ... end
10-76
Defining Abstract Classes

In this section... Abstract Classes on page 10-77 Declaring Classes as Abstract on page 10-78 Determine If a Class Is Abstract on page 10-79 Find Inherited Abstract Properties and Methods on page 10-80
Abstract Classes
Abstract classes are useful for describing functionality that is common to a group of classes, but requires unique implementations within each class. Abstract Class Terminology abstract class A class that cannot be instantiated, but that defines class components used by subclasses. abstract members Properties or methods declared in an abstract class, but implemented in subclasses. concrete members Properties or methods that are fully implement by a class. concrete class A class that can be instantiated. Concrete classes contain no abstract members. interface An abstract class describing functionality that is common to a group of classes, but that requires unique implementations within each class. The abstract class defines the interface of each subclass without specifying the actual implementation. An abstract class serves as a basis (that is, a superclass) for a group of related subclasses. An abstract class can define abstract properties and methods that subclasses must implement. Each subclass can implement the concrete properties and methods in a way that supports their specific requirements.
10-77
10
Abstract classes can define properties and methods that are not abstract, and do not need to define any abstract members. Abstract classes pass on their concrete members through inheritance.
Implementing a Concrete Subclass

A subclass must implement all inherited abstract properties and methods to become a concrete class. Otherwise, the subclass is itself an abstract class.
Declaring Classes as Abstract

A class is abstract when it declares: An abstract method An abstract property The Abstract class attribute A subclass of an abstract class is itself abstract if it does not define concrete implementations for all inherited abstract methods or properties.
Abstract Methods
Define an abstract method:
methods (Abstract) abstMethod(obj) end
For methods that declare the Abstract method attribute: Do not use a function...end block to define an abstract method, use only the method signature. Abstract methods have no implementation in the abstract class. Concrete subclasses are not required to support the same number of input and output arguments and do not need to use the same argument names. However, subclasses generally use the same signature when implementing their version of the method.
10-78
Abstract Properties
Define an abstract property:
properties (Abstract) AbsProp end
For properties that declare the Abstract property attribute: Concrete subclasses must redefine abstract properties without the Abstract attribute, and must use the same values for the SetAccess and GetAccess attributes as those used in the abstract superclass. Abstract properties cannot define set or get access methods (see Property Access Methods on page 6-14) and cannot specify initial values. The subclass that defines the concrete property can create set or get access methods and specify initial values.
Abstract Class
Declare a class as abstract in the classdef statement:
classdef (Abstract) AbsClass ... end
For classes that declare the Abstract class attribute: Concrete subclasses must redefine any properties or methods that are declared as abstract. The abstract class does not need to define any abstract methods or properties. When you define any abstract methods or properties, MATLAB automatically sets the class Abstract attribute to true.
Determine If a Class Is Abstract

Determine if a class is abstract by querying the Abstract property of its meta.class object. For example, the AbsClass defines two abstract methods:
10-79
10
classdef AbsClass methods(Abstract, Static) result = absMethodOne output = absMethodTwo end end
Use the logical value of the meta.class Abstract property to determine if the class is abstract:
mc = ?AbsClass; if ~mc.Abstract % not an abstract class end
Display Abstract Member Names

Use the meta.abstractDetails function to display the names of abstract properties or methods and the names of the defining classes:
meta.abstractDetails('AbsClass'); Abstract methods for class AbsClass: absMethodTwo % defined in AbsClass absMethodOne % defined in AbsClass
Find Inherited Abstract Properties and Methods

The meta.abstractDetails function returns the names and defining class of any inherited abstract properties or methods that you have not implemented in your subclass. This can be useful if you want the subclass to be concrete and need to determine what abstract members the subclass inherits. For example, suppose you subclass AbsClass, which is described in the Determine If a Class Is Abstract on page 10-79 section:
classdef SubAbsClass < AbsClass % Failed to implement absMethodOne % defined as abstract in AbsClass methods (Static) function out = absMethodTwo(a,b) out = a + b;
10-80
end end end
Determine if you implemented all inherited class members using meta.abstractDetails:

meta.abstractDetails(?SubAbsClass) Abstract methods for class SubAbsClass: absMethodOne % defined in AbsClass
The SubAbsClass class itself is abstract because it has not implemented the absMethodOne defined in AbsClass.
10-81
10
Defining Interfaces
In this section... Interfaces and Abstract Classes on page 10-82 An Interface for Classes Implementing Graphs on page 10-82
Interfaces and Abstract Classes

The properties and methods defined by a class form the interface that determines how class users interact with objects of the class. When creating a group of related classes, define a common interface to all these classes, even though the actual implementations of this interface can differ from one class to another. For example, consider a set of classes designed to represent various graphs (for example, line plots, bar graphs, pie charts, and so on). Suppose all classes must implement a Data property to contain the data used to generate the graph. However, the form of the data can differ considerably from one type of graph to another. Consequently, the way each class implements the Data property can be different. The same differences apply to methods. All classes can have a draw method that creates the graph, but the implementation of this method changes with the type of graph. The basic idea of an interface class is to specify the properties and methods that each subclass must implement without defining the actual implementation. This approach enables you to enforce a consistent interface to a group of related objects. As you add more classes in the future, the original interface remains.
An Interface for Classes Implementing Graphs

This example creates an interface for classes used to display specialized graphs. The interface is an abstract class that defines properties and methods that the subclasses must implement, but does not specify how to implement these components. This approach enforces the use of a consistent interface
10-82
Defining Interfaces
while providing the necessary flexibility to implement the internal workings of each specialized graph subclass differently. In this example, the interface, derived subclasses, and a utility function are contained in a package folder:
+graphics/graph.m +graphics/linegraph.m % abstract interface class % concrete subclass
Interface Properties and Methods

The graph class specifies the following properties, which the subclasses must define: Primitive Handle of the Handle Graphics object used to implement the specialized graph. The class user has no need to access these objects directly so this property has protected SetAccess and GetAccess. AxesHandle Handle of the axes used for the graph. The specialized graph objects can set axes object properties and also limit this propertys SetAccess and GetAccess to protected. Data All specialized graph objects must store data, but the type of data varies so each subclass defines the storage mechanism. Subclass users can change the data so this property has public access rights. The graph class names three abstract methods that subclasses must implement. The graph class also suggests in comments that each subclass constructor must accept the plot data and property name/property value pairs for all class properties. subclass_constructor Accept data and P/V pairs and return an object. draw Used to create a drawing primitive and render a graph of the data according to the type of graph implemented by the subclass. zoom Implementation of a zoom method by changing the axes CameraViewAngle property. The interface suggests the use of the camzoom function for consistency among subclasses. The zoom buttons created by the addButtons static method use this method as a callback.
10-83
10
updateGraph Method called by the set.Data method to update the plotted data whenever the Data property changes.
Interface Guides Class Design

The package of classes that derive from the graph abstract class implement the following behaviors: Creating an instance of a specialized graph object (subclass object) without rendering the plot Specifying any or none of the object properties when you create a specialized graph object Changing any object property automatically updates the currently displayed plot Allowing each specialized graph object to implement whatever additional properties it requires to give class users control over those characteristics.
Defining the Interface

The graph class is an abstract class that defines the methods and properties used by the subclasses. Comments in the abstract class suggest the intended implementation:
classdef graph < handle % Abstract class for creating data graphs % Subclass constructor should accept % the data that is to be plotted and % property name/property value pairs properties (SetAccess = protected, GetAccess = protected) Primitive % HG primitive handle AxesHandle % Axes handle end properties % Public access Data end methods (Abstract) draw(obj) % Use a line, surface, % or patch HG primitive
10-84
Defining Interfaces
zoom(obj,factor) % Change the CameraViewAngle % for 2D and 3D views % use camzoom for consistency updateGraph(obj) % Called by the set.Data method % to update the drawing primitive % whenever the Data property is changed end methods function set.Data(obj,newdata) obj.Data = newdata; updateGraph(obj) end function addButtons(gobj) hfig = get(gobj.AxesHandle,'Parent'); uicontrol(hfig,'Style','pushbutton','String','Zoom Out',... 'Callback',@(src,evnt)zoom(gobj,.5)); uicontrol(hfig,'Style','pushbutton','String','Zoom In',... 'Callback',@(src,evnt)zoom(gobj,2),... 'Position',[100 20 60 20]); end end end
The graph class implements the property set method (set.Data) to monitor changes to the Data property. An alternative is to define the Data property as Abstract and enable the subclasses to determine whether to implement a set access method for this property. However, by defining the set access method that calls an abstract method (updateGraph, which each subclass must implement), the graph interface imposes a specific design on the whole package of classes, without limiting flexibility.
Method to Work with All Subclasses

The addButtons method adds push buttons for the zoom methods, which each subclass must implement. Using a method instead of an ordinary function enables addButtons to access the protected class data (the axes handle). Use the objects zoom method as the push button callback.
function addButtons(gobj)
10-85
10
hfig = get(gobj.AxesHandle,'Parent'); uicontrol(hfig,'Style','pushbutton','String','Zoom Out',... 'Callback',@(src,evnt)zoom(gobj,.5)); uicontrol(hfig,'Style','pushbutton','String','Zoom In',... 'Callback',@(src,evnt)zoom(gobj,2),... 'Position',[100 20 60 20]); end
Deriving a Concrete Class linegraph

Note Display the fully commented code for the linegraph class by clicking this link: linegraph class. This example defines only a single subclass used to represent a simple line graph. It derives from graph, but provides implementations for the abstract methods draw, zoom, updateGraph, and its own constructor. The base class (graph) and subclass are all contained in a package (graphics), which you must use to reference the class name:
classdef linegraph < graphics.graph
Adding Properties
The linegraph class implements the interface defined in the graph class and adds two additional propertiesLineColor and LineType. This class defines initial values for each property, so specifying property values in the constructor is optional. You can create a linegraph object with no data, but you cannot produce a graph from that object.
properties LineColor = [0 0 0]; LineType = '-'; end
The linegraph Constructor

The constructor accepts a struct with x and y coordinate data, as well as property name/property value pairs:
10-86
Defining Interfaces
function gobj = linegraph(data,varargin) if nargin > 0 gobj.Data = data; if nargin > 2 for k=1:2:length(varargin) gobj.(varargin{k}) = varargin{k+1}; end end end end
Implementing the draw Method

The linegraph draw method uses property values to create a line object. The linegraph class stores the line handle as protected class data. To support the use of no input arguments for the class constructor, draw checks the Data property to determine if it is empty before proceeding:
function gobj = draw(gobj) if isempty(gobj.Data) error('The linegraph object contains no data') end h = line(gobj.Data.x,gobj.Data.y,... 'Color',gobj.LineColor,... 'LineStyle',gobj.LineType); gobj.Primitive = h; gobj.AxesHandle = get(h,'Parent'); end
Implementing the zoom Method

The linegraph zoom method follows the comments in the graph class which suggest using the camzoom function. camzoom provides a convenient interface to zooming and operates correctly with the push buttons created by the addButtons method.
Defining the Property Set Methods

Property set methods provide a convenient way to execute code automatically when the value of a property changes for the first time in a constructor. (See Property Set Methods on page 6-16.) The linegraph class uses set
10-87
10
methods to update the line primitive data (which causes a redraw of the plot) whenever a property value changes. The use of property set methods provides a way to update the data plot quickly without requiring a call to the draw method. The draw method updates the plot by resetting all values to match the current property values. Three properties use set methods: LineColor, LineType, and Data. LineColor and LineType are properties added by the linegraph class and are specific to the line primitive used by this class. Other subclasses can define different properties unique to their specialization (for example., FaceColor). The graph class implements the Data property set method. However, the graph class requires each subclass to define a method called updateGraph, which handles the update of plot data for the specific drawing primitive used.
Using the linegraph Class

The linegraph class defines the simple API specified by the graph base class and implements its specialized type of graph:
d.x = 1:10; d.y = rand(10,1); lg = graphics.linegraph(d,'LineColor','b','LineType',':'); lg.draw; graphics.graph.addButtons(lg);
Clicking the Zoom In button shows the zoom method providing the callback for the button.
10-88
Defining Interfaces
Changing properties updates the graph:

d.y = rand(10,1); % new set of random data for y lg.Data = d; lg.LineColor = [.9 .1 .6]; % LineColor can be char or double
Now click Zoom Out and see the new results:
10-89
10
10-90
11
Saving and Loading Objects
Understanding the Save and Load Process on page 11-2 Modifying the Save and Load Process on page 11-6 Maintaining Class Compatibility on page 11-9 Passing Arguments to Constructors During Load on page 11-14 Saving and Loading Objects from Class Hierarchies on page 11-17 Saving and Loading Dynamic Properties on page 11-20 Tips for Saving and Loading on page 11-22
11
Understanding the Save and Load Process

In this section... The Default Save and Load Process on page 11-2 When to Modify Object Saving and Loading on page 11-4
The Default Save and Load Process

Use save and load to store objects:
save filename object load filename object
What Information Is Saved

Saving objects in MAT-files saves: The full name of the objects class, including any package qualifiers. Values of dynamic properties. The names and current values of all properties, except:
Properties that have their Transient, Constant, or Dependent attributes set to true. See Specifying Property Attributes on page 6-7 for a description of property attributes.
Loading Property Data

When loading objects from MAT-files the load function: Creates a new object. Calls the class constructor with no arguments only if the classs ConstructOnLoad attribute is set to true. Assigns the saved values to the objects properties. These assignments results in calls to property set methods defined by the class. You can use property set methods to ensure property values are still valid in cases where the class definition has changed.
11-2
See Property Set Methods on page 6-16 for information on property set methods.
Errors During Load

It is possible for a default value to cause an error in a property set method (for example, the class definition might have changed). When an error occurs while an object is being loaded from a file, MATLAB returns the saved values in a struct. The field names correspond to the property names. In cases where the saved object is derived from multiple superclasses that define private properties having the same name, the struct contains the property value of the most direct superclass only.
Saving and Loading Deleted Handle Objects

If you save a deleted handle, MATLAB load it as a deleted handle. For example:
% Create a handle object >> a = containers.Map('Monday','sunny') isvalid(a) ans = 1 % Delete the handle object >> delete(a) >> isvalid(a) ans = 0 % Save the deleted handle >> save savefile a % Clear the variable a >> clear a % Load a back into the workspace
11-3
11
>> load savefile a >> isvalid(a) ans = 0
See the handle class delete method and the clear command for more information on these operations.
saveobj and loadobj

The save and load functions call your classs saveobj and loadobj methods, respectively, if your class defines these methods. You use these methods to customize the save and load process. When you issue a save command, MATLAB first calls your saveobj method and passes the output of saveobj to save. Similarly, when you call load, MATLAB passes the result of loading what you saved to loadobj. loadobj must then return a properly constructed object. Therefore, you must design saveobj and loadobj to work together.
When to Modify Object Saving and Loading

The following sections describe when and how to modify the process MATLAB uses to save and load objects. You modify this process by implementing saveobj and loadobj methods for your class.
Why Implement saveobj and loadobj

The primary reason for implementing saveobj and loadobj methods is to support backward and forward compatibility of classes. For example, you might have cases where: The classs properties have changed (just adding a new property does not necessarily require special code because it can be initialized to its default value when loaded). The order in which properties are initialized is important due to a circular reference to handle objects.
11-4
You must call the objects constructor with arguments and, therefore, cannot support a default constructor (no arguments).
Information to Consider
If you decide to modify the default save and load process, keep the following points in mind: If your loadobj method generates an error, MATLAB still loads the objects in whatever state the object was in before the invocation of loadobj. Subclass objects inherit superclass loadobj and saveobj methods. Therefore, if you do not implement a loadobj or saveobj method in the most specific class, MATLAB calls only the inherited methods. If a superclass implements a loadobj or saveobj method, then your subclass can also implement a loadobj or saveobj method that calls the superclass methods as necessary. See Saving and Loading Objects from Class Hierarchies on page 11-17 for more information. The load function does not call the default constructor by default. See Calling Constructor When Loading on page 11-25 for more information. If an error occurs while the object is loading from a file, the load function passes your loadobj method as much data as it can successfully load from the file. In case of an error, load passes loadobj a struct whose field names correspond to the property names extracted from the file. See Reconstructing Objects with loadobj on page 11-15 for an example of a loadobj method that processes a struct. See Tips for Saving and Loading on page 11-22 for guidelines on saving and loading objects.
11-5
11
Modifying the Save and Load Process

In this section... Class saveobj and loadobj Methods on page 11-6 Processing Objects During Load on page 11-7 Save and Load Applications on page 11-7
Class saveobj and loadobj Methods

You can define methods for your class that are executed when you call save or load on an object: The save function calls your classs saveobj method before performing the save operation. The save function then saves the value returned by the objects saveobj method. You can use the saveobj method to return a modified object or any other type of variable, such as a struct array. The load function calls your classs loadobj method after loading the object. The load function loads into the workspace the value returned by the objects loadobj method. If you define a loadobj method you can modify the object being returned or reconstruct an object from the data saved by your saveobj method. If you implement a saveobj method that modifies the object being saved, implement a loadobj method to return the object to its proper state when reloading it. For example, you might want to store an objects data in a struct array and reconstruct the object when reloaded to manage changes to the class definition.
Implement loadobj as a Static Method

You must implement the loadobj method as a Static method because loadobj can actually be called with a struct or other data instead of an object of the class. You can implement the saveobj method as an ordinary method (i.e., calling it requires an instance of the class). MATLAB saves the objects class name so that load can determine which loadobj method to call, even if your saveobj method saves only the objects data in an array and not the object itself.
11-6
Modifying the Save and Load Process
Processing Objects During Load

Implementing a loadobj method enables you to apply some processing to the object before it is loaded into the workspace. You might need to do this if: The class definition has changed since the object was saved and you need to modify the object before reloading. A saveobj method modified the object during the save operation, perhaps saving data in an array, and the loadobj method must reconstruct the object based on the output of saveobj.
Updating an Object Property When Loading

In the following example, the loadobj method checks if the object to be loaded has an old, shorter account number and calls a function to return an updated account number if necessary. After updating the objects AccountNumber property, loadobj returns the object to be loaded into the workspace.
methods (Static = true) function obj = loadobj(a) accnb = a.AccountNumber; if length(num2str(accnb)) < 12 a.AccountNumber = updateAccountNumber(accnb); % update object end obj = a; end end % return the updated object
In this case, you do not need to implement a saveobj method. You are using loadobj only to ensure older saved objects are brought up to date before loading. The Save and Load Applications on page 11-7 section provides an example in which loadobj performs specific operations to recreate an object based on the data returned by saveobj during the save operation.
Save and Load Applications

The following sections describe some specific applications involving the saving and loading of objects.
11-7
11
Maintaining Class Compatibility on page 11-9 how to maintain compatibility among progressive versions of an application. Passing Arguments to Constructors During Load on page 11-14 using loadobj to call the class constructor of an object when you need to pass arguments to the constructor during load. Saving and Loading Objects from Class Hierarchies on page 11-17 how inherited methods affect saving and loading objects. Saving and Loading Dynamic Properties on page 11-20 how to handle dynamic properties when saving and loading objects.
11-8
Maintaining Class Compatibility

Versions of a Phone Book Application Program
This section shows you how to use saveobj and loadobj methods to maintain compatibility among subsequent releases of an application program. Suppose you have created a program that implements a phone book application, which can be used to keep track of information about various people and companies. One of the key elements of this program is that it uses a data structure to contain the information for each phone book entry. You save these data structures in MAT-files. This example shows ways to maintain the compatibility of subsequent versions of the data structures as you implement new versions of the program. When the phone book application program loads a particular phone book entry by reading a variable from a Mat-file, it must ensure that the loaded data can be used by the current version of the application.
Version 1 Stores Data in struct

Suppose in Version 1 of the phone book application program, you used an ordinary MATLAB struct to save phone book entries in the fields: Name, Address, and PhoneNumber. Your phone book application program saves these variables in a MAT-file. For example, here is a typical entry:
V1.Name = 'The MathWorks, Inc.'; V1.Address = '3 Apple Hill Drive, Natick, MA, 01760'; V1.PhoneNumber = '5086477000';
Version 2 Maps struct Fields to Object Properties

With Version 2 of the phone book program, you change from a struct to a class having public properties with the same names as the fields in the struct. You want to save the new PhoneBookEntry objects and you want to load the old struct without causing any errors. To maintain this compatibility, the PhoneBookEntry class implements loadobj and saveobj methods:
classdef PhoneBookEntry properties Name
11-9
11
Address PhoneNumber end methods (Static) function obj = loadobj(obj) if isstruct(obj) % Call default constructor newObj = PhoneBookEntry; % Assign property values from struct newObj.Name = obj.Name; newObj.Address = obj.Address; newObj.PhoneNumber = obj.PhoneNumber; obj = newObj; end end end methods function obj = saveobj(obj) s.Name = obj.Name; s.Address = obj.Address; s.PhoneNumber = obj.PhoneNumber; obj = s; end end end saveobj saves the object data in a struct that uses property names for field names. This struct is compatible with Version 1 of the product. When the struct is loaded into Version 2 of the phone book application program, the static loadobj method converts the struct to a PhoneBookEntry object. For example, given the previously defined struct V1: V1 = Name: 'MathWorks, Inc.' Address: '3 Apple Hill Drive, Natick, MA, 01760' PhoneNumber: '5086477000'
The application program can use the loadobj static method to convert this Version 1 struct to a Version 2 object:
11-10
V2 = PhoneBookEntry.loadobj(V1) V2 = PhoneBookEntry with properties: Name: 'MathWorks, Inc.' Address: '3 Apple Hill Drive, Natick, MA, 01760' PhoneNumber: '5086477000'
If a Version 2 PhoneBookEntry object is loaded, load automatically calls the objects loadobj method, which converts the struct to an object compatible with Version 2 of the phone book application program.
Version 3 Adds More Properties to Class

In Version 3, you change the PhoneBookEntry class by splitting the Address property into StreetAddress, City, State, and ZipCode properties. With this version, you cannot load a Version 3 PhoneBookEntry object in previous releases by default. However, the saveobj method provides an option to save Version 3 objects as structs that you can load in Version 2. The loadobj method enables you to load both Version 3 objects and Version 2 structs. Here is the new version of the PhoneBookEntry class.
classdef PhoneBookEntry properties Name StreetAddress City State ZipCode PhoneNumber end properties (Constant) Sep = ', '; end properties (Dependent, SetAccess=private) Address end properties (Transient)
11-11
11
SaveInOldFormat = 0; end methods (Static) function obj = loadobj(obj) if isstruct(obj) % Call default constructor newObj = PhoneBookEntry; % Assign property values from struct newObj.Name = obj.Name; newObj.Address = obj.Address; newObj.PhoneNumber = obj.PhoneNumber; obj = newObj; end end end methods function address = get.Address(obj) address=[obj.StreetAddress obj.Sep obj.City obj.Sep obj.State obj.Sep obj.ZipCode]; end function obj = set.Address(obj,address) addressItems = regexp(address,obj.Sep,'split'); if length(addressItems) == 4 obj.StreetAddress = addressItems{1}; obj.City = addressItems{2}; obj.State = addressItems{3}; obj.ZipCode = addressItems{4}; else error('PhoneBookEntry:InvalidAddressFormat', ... 'Invalid address format.'); end end function obj = saveobj(obj) % If set to true, save as a struct if obj.SaveInOldFormat s.Name = obj.Name; s.Address = obj.Address; s.PhoneNumber = obj.PhoneNumber; obj = s; end end
11-12
end
To maintain compatibility among all versions, Version 3 of the PhoneBookEntry class applies the following techniques: Preserve the Address property (which is used in Version 2) as a Dependent property with private SetAccess. Define an Address property get method (get.Address) to build a string that is compatible with the Version 2 Address property. The get.Address method is invoked from the saveobj method to assign the object data to a struct that is compatible with previous versions. The struct continues to have only an Address field built from the data in the new StreetAddress, City, State, and ZipCode properties. As the loadobj method sets the objects Address property, it invokes the property set method (set.Address), which extracts the substrings required by the StreetAddress, City, State, and ZipCode properties. The Transient (not saved) property SaveInOldFormat enables you to specify whether to save the Version 3 object as a struct or an object. See Property Access Methods on page 6-14 for more on property set and get methods.
11-13
11
Passing Arguments to Constructors During Load

In this section... Calling Constructors When Loading Objects on page 11-14 Code for This Example on page 11-14 Example Overview on page 11-14
Calling Constructors When Loading Objects

You can set the class ConstructOnLoad attribute when you need to call the default (no argument) class constructor on an object that is loaded from a MAT-file. Then load automatically calls the objects class constructor, but cannot pass any arguments to it. If the object you are loading requires a call to its class constructor and this call requires you to pass arguments to the constructor, you can implement a loadobj method that performs this task. For example, suppose the objects constructor adds a listener and, therefore, must be passed a handle to the object triggering the event (required by the addlistener handle class method) to create this listener. Your loadobj method could call the constructor with the required argument.
Code for This Example

The following information on saving and loading objects refers to a BankAccountSL class. Click the following link to open the full code for this class in the MATLAB editor: Open class definition in editor
Example Overview
This example shows how to use loadobj to call a class constructor with arguments at load time. Because the constructor requires arguments, you cannot use the ConstructOnLoad attribute to load the object, which causes a call to the default (no arguments) constructor.
11-14
Passing Arguments to Constructors During Load
This example uses loadobj to determine the status of a BankAccountSL object when the object data is loaded, and then calls the class constructor with the appropriate arguments to create the object. This approach provides a way to modify the criteria for determining status over time, while ensuring that all loaded objects are using the current criteria. The saveobj method extracts the data from the object and writes this data into a struct, which saveobj returns to the save function.
Saving Only Object Data with saveobj

The following saveobj method saves the values of the BankAccountSL objects AccountNumber and AccountBalance properties in the struct variable A, which has field names that match the property names. saveobj then returns the variable A to be saved in the MAT-file by the save function.
methods function A = saveobj(obj) A.AccountNumber = obj.AccountNumber; A.AccountBalance = obj.AccountBalance; end end
Reconstructing Objects with loadobj

The BankAccountSL class AccountStatus property is Transient because its value depends on the value of the AccountBalance property and the current criteria and possible status values. You can use the loadobj method to update all saved BankAccount objects when they are loaded into your system. To create a valid object, loadobj calls the constructor using the data saved in the struct A and passes any other required arguments. If the account balance is greater than zero, AccountStatus is set to open. If the account balance is zero or less, AccountStatus is set to overdrawn or to frozen. The following loadobj method calls the class constructor with the appropriate values for the arguments:
methods (Static)
11-15
11
function obj = loadobj(A) if A.AccountBalance > 0 obj = BankAccountSL(A.AccountNumber,A.AccountBalance,'open'); elseif else obj = BankAccountSL(A.AccountNumber,A.AccountBalance,'frozen'); end end end A.AccountBalance < 0) && (A.AccountBalance >= -100) obj = BankAccountSL(A.AccountNumber,A.AccountBalance,'overdrawn');
11-16
Saving and Loading Objects from Class Hierarchies

Saving and Loading Subclass Objects
When you modify the save operation of an object that is part of a class hierarchy, you must be sure that all classes in the hierarchy perform the correct operations in the save and load process. If the most specific class of an object does not define a loadobj or saveobj method, this class can inherit loadobj or saveobj methods from a superclass. If any class in the hierarchy defines special save and load behavior: Define saveobj for all classes in the hierarchy. Call superclass saveobj methods from the subclass saveobj method because the save function calls only one saveobj method. If saveobj returns a struct instead of the object, then the subclass can implement a loadobj method to reconstruct the object. The subclass loadobj method can call the superclass loadobj, or other methods as required, to assign values to their properties.
Reconstructing the Subclass Object from a Saved Struct

Suppose you want to save a subclass object by first converting its property data to a struct in the classs saveobj method and then reconstruct the object when loaded using its loadobj method. This action requires that: Superclasses implement saveobj methods to save their property data in the struct. The subclass saveobj method calls each superclass saveobj method and then returns the completed struct to the save function, which writes the struct to the MAT-file. The subclass loadobj method creates a subclass object and then calls superclass methods to assign their property values in the subclass object. The subclass loadobj method returns the reconstructed object to the load function, which loads the object into the workspace.
11-17
11
The following superclass (MySuper) and subclass (MySub) definitions show how to code these methods. The MySuper class defines a loadobj method to enable an object of this class to be loaded directly. The subclass loadobj method calls a method named reload after it constructs the subclass object. reload first calls the superclass reload method to assign superclass property values and then assigns the subclass property value.
classdef MySuper % Superclass definition properties X Y end methods function S = saveobj(obj) % Save property values in struct % Return struct for save function to write to MAT-file S.PointX = obj.X; S.PointY = obj.Y; end function obj = reload(obj,S) % Method used to assign values from struct to properties % Called by loadobj and subclass obj.X = S.PointX; obj.Y = S.PointY; end end methods (Static) function obj = loadobj(S) % Constructs a MySuper object % loadobj used when a superclass object is saved directly % Calls reload to assign property values retrived from struct % loadobj must be Static so it can be called without object obj = MySuper; obj = reload(obj,S); end end end
11-18
Your subclass implements saveobj and loadobj methods that call superclass methods.
classdef MySub < MySuper % Subclass definition properties Z end methods function S = saveobj(obj) % Call superclass saveobj % Save property values in struct S = saveobj@MySuper(obj); S.PointZ = obj.Z; end function obj = reload(obj,S) % Call superclass reload method % Assign subclass property value % Called by loadobj obj = reload@MySuper(obj,S); obj.Z = S.PointZ; end end methods (Static) function obj = loadobj(S) % Create object of MySub class % Assign property value retrived from struct % loadobj must be Static so it can be called without object obj = MySub; obj = reload(obj,S); end end end
11-19
11
Saving and Loading Dynamic Properties

Reconstructing Objects That Have Dynamic Properties
If you use the addprop method to add dynamic properties to a MATLAB class derived from the dynamicprops class, those dynamic properties are saved along with the object to which they are attached when you save the object to a MAT-file. See Dynamic Properties Adding Properties to an Instance on page 6-27 for more information about dynamic properties.
Why You Need saveobj and loadobj Methods

save saves dynamic properties and their values. However, save does not save dynamic property attributes because these attributes are not specified in the class definition. If you are saving an object that has dynamic properties, and these properties use nondefault attributes, you need to manage the saving and loading of attribute values using saveobj and loadobj.
If your class implements a saveobj method that converts the object to another type of MATLAB variable, such as a struct, you can save the dynamic propertys attribute values so that your loadobj method can reconstruct these properties. The attribute values of dynamic properties are not part of the class definition and might have been set after the properties were attached to the object, so these values might not be known to the loadobj method.
Implementing the saveobj and loadobj Methods

For example, your saveobj method can obtain the nondefault attribute values from the dynamic propertys meta.DynamicProperty. Suppose the object you are saving has a dynamic property called DynoProp, and your saveobj method creates a struct s to save the data that the loadobj method uses to reconstruct the object:
methods function s = saveobj(obj) ... % Obtain the meta.DynamicProperty object for the dynamic property metaDynoProp = findprop(obj,'DynoProp'); % Record name and value for the dynamic property s.dynamicprops(1).name = metaDynoProp.Name;
11-20
Saving and Loading Dynamic Properties
s.dynamicprops(1).value = obj.DynoProp; % Record additional dynamic property attributes so they can be % restored at load time, for example SetAccess and GetAccess s.dynamicprops(1).setAccess = metaDynoProp.SetAccess; s.dynamicprops(1).getAccess = metaDynoProp.GetAccess; ... end end
Your loadobj method can add the dynamic property and set the attribute values:
methods (Static) function obj = loadobj(s) % first, create an instance of the class obj = ClassConstructor; ... % Add new dynamic property to object metaDynoProp = addprop(obj,s.dynamicprops(1).name); obj.(s.dynamicprops(1).name) = s.dynamicprops(1).value; % Restore dynamic property attributes metaDynoProp.SetAccess = s.dynamicprops(1).setAccess; metaDynoProp.GetAccess = s.dynamicprops(1).getAccess; end end
11-21
11
Tips for Saving and Loading

In this section... Using Default Property Values to Reduce Storage on page 11-22 Avoiding Property Initialization Order Dependency on page 11-23 When to Use Transient Properties on page 11-25 Calling Constructor When Loading on page 11-25
Using Default Property Values to Reduce Storage

When loading an object, MATLAB creates a new object and assigns the stored property values. For properties that had default values at the time you saved the object, MATLAB loads the saved default values, even if the class definition defines new default values for those properties. See Defining Default Values on page 3-11 for more information on how MATLAB evaluates default value expressions.
Reducing Object Storage

If a property is often set to the same value, define a default value for that property. When the object is saved to a MAT-file, MATLAB does not save the default value, thereby, saving storage space.
Implementing Forward and Backward Compatibility

Default property values can help you implement version compatibility for saved objects. For example, if you add a new property to version 2 of your class, having a default value enables MATLAB to assign a value to the new property when loading a version 1 object. Similarly, if version 2 of your class removes a property, then if a version 2 object is saved and loaded into version 1, your loadobj method can use the default value from version 1 for the version 2 object.
11-22
Avoiding Property Initialization Order Dependency

Use a Dependent property when the property value needs to be calculated at runtime. Whenever you can use a dependent property in your class definition you save storage for saved objects. Dependent is a property attribute (see Property Attributes on page 6-8 for a complete list.)
Controlling Property Loading

If your class design is such that setting one property value causes other property values to be updated, then you can use dependent properties to ensure objects load properly. For example, consider the following Odometer class. It defines two public properties: TotalDistance and Units. Whenever Units is modified, the TotalDistance is modified to reflect the change. There is also a private property, PrivateUnits, and a constant property ConversionFactor.
classdef Odometer properties(Constant) ConversionFactor = 1.6 end properties TotalDistance = 0 end properties(Dependent) Units end properties(Access=private) PrivateUnits = 'mi' end methods function unit = get.Units(obj) unit = obj.PrivateUnits; end function obj = set.Units(obj, newUnits) % validate newUnits to be a string switch(newUnits) case 'mi' if strcmp(obj.Units, 'km') obj.TotalDistance = obj.TotalDistance / ... obj.ConversionFactor;
11-23
11
obj.PrivateUnits = newUnits; end case 'km' if strcmp(obj.Units, 'mi') obj.TotalDistance = obj.TotalDistance * ... obj.ConversionFactor; obj.PrivateUnits = newUnits; end otherwise error('Odometer:InvalidUnits', ... 'Units ''%s'' is not supported.', newUnits); end end end end
Suppose you create an instance of Odometer with the following property values:
odObj = Odometer; odObj.Units = 'km'; odObj.TotalDistance = 16;
When you save the object, the following happens to property values: ConversionFactor is not saved because it is a Constant property. TotalDistance is saved. Units is not saved because it is a Dependent property. PrivateUnits is saved and provides the storage for the current value of Units. When you load the object, the following happens to property values: ConversionFactor is obtained from the class definition. TotalDistance is loaded from the saved object. Units is not loaded so its set method is not called.
11-24
PrivateUnits is loaded and contains the value that is used if the Units get method is called. If the Units property was not Dependent, loading it calls its set method and causes the TotalDistance property to be set again.
When to Use Transient Properties

The value of a Transient property is never stored when an object is saved to a file, but instances of the class do allocate storage to hold a value for this property. These two characteristics make a Transient property useful for cases where data needs to be stored in the object temporarily as an intermediate computation step, or for faster retrieval. (See Property Attributes on page 6-8 for a complete list of properties.) You can use Transient properties to reduce storage space and simplify the load process in cases where: The property data can be easily reproduced at run-time. The property represent intermediate state that you can discard
Calling Constructor When Loading

MATLAB does not call the class constructor when loading an object from a MAT-file. However, if you set the ConstructOnLoad class attribute to true, load does call the constructor with no arguments. Enabling ConstructOnLoad is useful when you do not want to implement a loadobj method, but do need to perform some actions at construction time, such as registering listeners for another object. You must be sure that the class constructor can be called with no arguments without generating an error. See No Input Argument Constructor Requirement on page 7-19. In cases where the class constructor sets only some property values based on input arguments, then using ConstructOnLoad is probably not useful. See Passing Arguments to Constructors During Load on page 11-14 for an alternative.
11-25
11
11-26
12
Enumerations
Defining Named Values on page 12-2 Working with Enumerations on page 12-4 Enumerations Derived from Built-In Types on page 12-16 Mutable (Handle) vs. Immutable (Value) Enumeration Members on page 12-22 Enumerations That Encapsulate Data on page 12-30 Saving and Loading Enumerations on page 12-35
12
Enumerations
Defining Named Values

Kinds of Predefined Names
MATLAB supports two kinds of predefined names: Constant properties Enumerations
Constant Properties
Use constant properties when you want a collection of related constant values whose values can belong to different types (numeric values, character strings, and so on). Define properties with constant values by setting the property Constant attribute. Reference constant properties by name whenever you need access to that particular value. See Properties with Constant Values on page 13-2 for more information.
Enumerations
Use enumerations when you want to create a fixed set of names representing a single type of value. You can derive enumeration classes from other classes to inherit the operations of the superclass. For example, if you define an enumeration class that subclasses a MATLAB numeric class like double or int32, the enumeration class inherits all of the mathematical and relational operations that MATLAB defines for those classes. Using enumerations instead of character strings to represent a value, such as colors ('red'), can result in more readable code because: You can compare enumeration members with == instead of using strcmp Enumerations maintain type information, strings do not. For example, passing a string 'red' to functions means that every function must interpret what 'red' means. If you define red as an enumeration, the actual value of 'red' can change (from [1 0 0] to [.93 .14 .14], for example) without updating every function that accepts colors, as you would if you defined the color as a string 'red'.
12-2
Defining Named Values
Define enumerations by creating an enumeration block in the class definition. See Working with Enumerations on page 12-4 for more information.
12-3
12
Enumerations
Working with Enumerations

In this section... Basic Knowledge on page 12-4 Using Enumeration Classes on page 12-5 Defining Methods in Enumeration Classes on page 12-9 Defining Properties in Enumeration Classes on page 12-9 Array Expansion Operations on page 12-11 Constructor Calling Sequence on page 12-11 Restrictions Applied to Enumeration Classes on page 12-13 Techniques for Defining Enumerations on page 12-13
Basic Knowledge
The material presented in this section builds on an understanding of the information provided in the following sections. Defining Classes and Class Members Class Syntax Fundamentals Creating Subclasses Syntax and Techniques on page 10-7 Mutable and Immutable Properties on page 6-13 enumeration function displays enumeration names Terminology and Concepts This documentation uses terminology as described in the following list: Enumeration or Enumeration class A class that contains an enumeration block defining enumeration members. Enumeration member A named instance of an enumeration class. Enumeration member constructor arguments Values in parentheses next to the enumeration member name in the enumeration block. When you
12-4
create an instance of an enumeration member, MATLAB passes the value or values in parenthesis to the class constructor. Underlying value For enumerations derived from built-in classes, the value associated with an instance of an enumeration class (that is, an enumeration member).
Using Enumeration Classes

Create an enumeration class by adding an enumeration block to a class definition. For example, the WeekDays class enumerates a set of days of the week.
classdef WeekDays enumeration Monday, Tuesday, Wednesday, Thursday, Friday end end
Constructing an Enumeration Member

Refer to an enumeration member using the class name and the member name:
ClassName.MemberName
For example, assign the enumeration member WeekDays.Tuesday to the variable today:
today = WeekDays.Tuesday; today is a variable of class WeekDays: >> whos Name today >> today today = Tuesday
Size 1x1
Bytes 56
Class WeekDays
Attributes
12-5
12
Enumerations
Default Methods
Enumeration classes have four methods by default:
>> methods(today) Methods for class WeekDays: WeekDays char eq ne
Default constructor (WeekDays in this case) char converts enumeration members to character strings eq enables use of == in expressions ne enables use of ~= in expressions Equality and inequality methods enable you to use enumeration members in if and switch statements and other functions that test for equality. Because you can define enumeration members with descriptive names, conversion to char is useful. For example:
today = WeekDays.Friday; ['Today is ',char(today)] ans = Today is Friday
Testing for Membership in a Set

Suppose you want to determine if today is a meeting day for your team.
today = WeekDays.Tuesday; teamMeetings = [WeekDays.Wednesday WeekDays.Friday];
Use the ismember function to determine if today is a meeting day:

ismember(today,teamMeetings) ans = 0
12-6
Using Enumerations in a Switch Statement

Enumerations work in switch statements:
function c = Reminder(day) % Add error checking here switch(day) case WeekDays.Monday c = 'Department meeting at 10:00'; case WeekDays.Tuesday c = 'Meeting Free Day!'; case {WeekDays.Wednesday WeekDays.Friday} c = 'Team meeting at 2:00'; case WeekDays.Thursday c = 'Volley ball night'; end end
Pass a member of the WeekDays enumeration class to the Reminder function:

>> today = WeekDays.Wednesday; >> Reminder (today) ans = Team meeting at 2:00
See Objects In Switch Statements on page 3-34 for more information.
Getting Information About Enumerations

You can get information about enumeration classes using the enumeration function. For example:
enumeration WeekDays Enumeration members for class 'WeekDays': Monday Tuesday Wednesday Thursday
12-7
12
Enumerations
Friday
See also Metaclass EnumeratedValues Property on page 14-7
Converting to Superclass Value

If an enumeration class specifies a superclass, in many cases you can convert an enumeration object to the superclass by passing the object to the superclass constructor. However, the superclass must be able to accept its own class as input and return an instance of the superclass. MATLAB built-in numeric classes, like double, single, and so on allow this conversion. For example, the Bearing class derives from the uint32 built-in class:
classdef Bearing < uint32 enumeration North (0) East (90) South (180) West (270) end end
Assign the Bearing.East member to the variable a:

a = Bearing.East;
Pass a to the superclass constructor and return an object of the superclass, b:

b = uint32(a); whos Name Size a b 1x1 1x1
Bytes 60 4
Class Bearing uint32
Attributes
The uint32 constructor accepts an instance of the subclass Bearing and returns and object of class uint32.
12-8
Defining Methods in Enumeration Classes

Define methods in an enumeration class like any MATLAB class. For example, here is the WeekDays class with a method called isMeetingDay added:
classdef WeekDays enumeration Monday, Tuesday, Wednesday, Thursday, Friday end methods function tf = isMeetingDay(obj) tf = ~(WeekDays.Tuesday == obj); end end end
Call isMeetingDay with an instance of the WeekDays class:

>> today = WeekDays.Tuesday; >> today.isMeetingDay ans = 0
You can pass the enumeration member to the method directly:

>> isMeetingDay(WeekDays.Wednesday) ans = 1
Defining Properties in Enumeration Classes

Add properties to an enumeration class when you must store data related to the enumeration members. Set the property values in the class constructor. For example, the SyntaxColors class defines three properties whose values the constructor assigns to the values of the input arguments when you reference a class member.
classdef SyntaxColors properties
12-9
12
Enumerations
R G B end methods function c = SyntaxColors(r, g, b) c.R = r; c.G = g; c.B = b; end end enumeration Error (1, 0, 0) Comment (0, 1, 0) Keyword (0, 0, 1) String (1, 0, 1) end end
When you refer to an enumeration member, the constructor initializes the property values:
e = SyntaxColors.Error; e.R ans = 1
Because SyntaxColors is a value class (it does not derive from handle), only the class constructor can set property values:
e.R = 0 Setting the 'R' property of the 'SyntaxColors' class is not allowed.
See Mutable (Handle) vs. Immutable (Value) Enumeration Members on page 12-22 for more information on enumeration classes that define properties.
12-10
Array Expansion Operations

MATLAB enables assignment to any element of an array, even if the array does not exist. For example, you can create an array of WeekDays objects:
classdef WeekDays enumeration Monday, Tuesday, Wednesday, Thursday, Friday end end clear ary(5) = WeekDays.Tuesday;
MATLAB must initialize the values of array elements ary(1:4). The default value of an enumeration class is the first enumeration member defined by the class in the enumeration block. The result of the assignment to the fifth element of the array ary is, therefore:
ary ary = Monday Monday Monday Monday Tuesday
Constructor Calling Sequence

Each statement in an enumeration block is the name of an enumeration member, optionally followed by an argument list. If the enumeration class defines a constructor, MATLAB calls the constructor to create the enumerated instances. MATLAB provides a default constructor for all enumeration classes that do not explicitly define a constructor. The default constructor creates an instance of the enumeration class: Using no input arguments, if the enumeration member defines no input arguments Using the input arguments defined in the enumeration class for that member For example, the input arguments for the Boolean class are 0 for Boolean.No and 1 for Boolean.Yes.
12-11
12
Enumerations
classdef Boolean < logical enumeration No (0) Yes (1) end end
The values of 0 and 1 are of class logical because the default constructor passes the argument to the first superclass. That is,
n = Boolean.No;
results in a call to logical that is equivalent to the following statement in a constructor:

function obj = Boolean(val) obj@logical(val) end
MATLAB passes the member argument only to the first superclass. For example, suppose Boolean derived from another class:
classdef Boolean < logical & MyBool enumeration No (0) Yes (1) end end
The MyBool class can add some specialized behavior:

classdef MyBool methods function boolValues = testBools(obj) ... end end end
Now, the default Boolean constructor behaves as if defined like this function:
function obj = Boolean(val)
12-12
obj@logical(val) % Argument passed to first superclass constructor obj@MyBool % No arguments passed to subsequent constructors end
Restrictions Applied to Enumeration Classes

Enumeration classes, which consist of a fixed set of possible values, restrict certain aspects of class use and definition: Enumeration classes are implicitly Sealed. You cannot define a subclass of an enumeration class because doing so would expand the set. You cannot call the constructor of an enumeration class directly. Only MATLAB can call enumeration class constructors to create the fixed set of members. The properties of value-based enumeration classes are immutable. Only the constructor can assign property values. MATLAB implicitly defines the SetAccess attributes of all properties defined by value-based enumeration classes as immutable. You cannot set the SetAccess attribute to any other value. All properties inherited by a value-based enumeration class that are not defined as Constant must have immutable SetAccess. The properties of handle-based enumeration classes are mutable. You can set property values on instances of the enumeration class. See Mutable (Handle) vs. Immutable (Value) Enumeration Members on page 12-22 for more information. An enumeration member cannot have the same name as a property, method, or event defined by the same class. Enumerations do not support colon (a:b) operations. For example, FlowRate.Low:FlowRate.High causes an error even if the FlowRate class derives from a numeric superclass.
Techniques for Defining Enumerations

Enumerations enable you to define names that represent entities useful to your application, without using numeric values or character strings. All enumerations support equality and inequality operations. Therefore, switch, if, and a number of comparison functions like isequal and ismember work with enumeration members.
12-13
12
Enumerations
You can define enumeration classes in ways that are most useful to your application, as described in the following sections.
Simple Enumerated Names

Simple enumeration classes have no superclasses and no properties. These classes define a set of related names that have no underlying values associated with them. Use this kind of enumeration when you want descriptive names, but your application does not require specific information associated with the name. See the WeekDays class in the Using Enumeration Classes on page 12-5 and the Defining Methods in Enumeration Classes on page 12-9 sections.
Enumerations with Built-In Class Behaviors

Enumeration classes that subclass MATLAB built-in classes inherit most of the behaviors of those classes. For example, an enumeration class derived from the double class inherits the mathematical, relational, and set operations that work with variables of the class. Enumerations do not support the colon (:) operator, even if the superclass does. See Restrictions Applied to Enumeration Classes on page 12-13 for more information. See Enumerations Derived from Built-In Types on page 12-16.
Enumerations with Properties for Member Data

Enumeration classes that do not subclass MATLAB built-in numeric and logical classes can define properties. These classes can define constructors that set each members unique property values. The constructor can save input arguments in property values. For example, a Color class can specify a Red enumeration member color with three (Red, Green, Blue) values:
enumeration Red (1,0,0) end
12-14
See Enumerations That Encapsulate Data on page 12-30
12-15
12
Enumerations

In this section... Basic Knowledge on page 12-16 Why Derive Enumerations from Built-In Types on page 12-16 Aliasing Enumeration Names on page 12-18 Superclass Constructor Returns Underlying Value on page 12-19 Default Converter on page 12-20
Basic Knowledge
The material presented in this section builds on an understanding of the information provided in the following sections. Fundamental MATLAB Classes for information on MATLAB built-in classes. enumeration function displays enumeration names
Why Derive Enumerations from Built-In Types

Note Enumeration classes derived from built-in numeric and logical classes cannot define properties. If an enumeration class subclasses a built-in numeric class, the subclass inherits ordering and arithmetic operations, which you can apply to the enumerated names. For example, the Results class subclasses the int32 built-in class and associates an integer value with each of the four enumeration members First, Second, Third, and NoPoints.
classdef Results < int32 enumeration First (100) Second (50)
12-16
Third (10) NoPoints (0) end end
Because the enumeration member inherits the methods of the int32 class (not the colon operator), you can use these enumerations like numeric values (summed, sorted, averaged, and so on).
isa(Results.Second,'int32') ans = 1
For example, use enumeration names instead of numbers to rank two teams:
Team1 = [Results.First, Results.NoPoints, Results.Third, Results.Second]; Team2 = [Results.Second, Results.Third, Results.First, Results.First];
Perform int32 operations on these Results enumerations:

sum(Team1) ans = 160 mean(Team1) ans = 40 sort(Team2,'descend') ans = First Team1 > Team2 ans = 1 0 0 0 sum(Team1) < sum(Team2) First Second Third
12-17
12
Enumerations
ans = 1
Creating Enumeration Instances

When you first refer to an enumeration class that derives from a built-in class such as, int32, MATLAB passes the input arguments associated with the enumeration members to the superclass constructor. For example, referencing the Second Results member, defined as:
Second (50)
means that MATLAB calls:

int32(50)
to initialize the int32 aspect of this Results object.
Aliasing Enumeration Names

Enumeration classes that derive from MATLAB built-in numeric and logical classes can define more than one name for an underlying value. The first name in the enumeration block with a given underlying value is the actual name for that underlying value and subsequent names are aliases. Specify aliased names with the same superclass constructor argument as the actual name:
classdef Boolean < logical enumeration No (0) Yes (1) off (0) on (1) end end
For example, the actual name of an instance of the Boolean.off enumeration member is No:
12-18
a = Boolean.No a = No b = Boolean.off b = No
Superclass Constructor Returns Underlying Value

The actual underlying value associated with an enumeration member is the value returned by the built-in superclass. For example, consider the Boolean class defined with constructor arguments that are of class double:
classdef Boolean < logical enumeration No (0) Yes (100) end end
This class derives from the built-in logical class. Therefore, underlying values for an enumeration member depend only on what value logical returns when passed that value:
a = Boolean.Yes a = Yes logical(a) ans = 1
12-19
12
Enumerations
Subclassing a Numeric Built-In Class

The FlowRate enumeration class defines three members, Low, Medium, and High.
classdef FlowRate < int32 enumeration Low (10) Medium (50) High (100) end end
Referencing an instance of an enumeration member:

setFlow = FlowRate.Medium;
returns an instance that is the result of MATLAB calling the default constructor with the argument value of 50. MATLAB passes this argument to the first superclass constructor (int32(50) in this case), which results in an underlying value of 50 as a 32-bit integer for the FlowRate.Medium member. Because FlowRate subclasses a MATLAB built-in numeric class (int32), it cannot define properties. However FlowRate inherits int32 methods including a converter method, which programs can use to obtain the underlying value:
setFlow = FlowRate.Medium; int32(setFlow) ans = 50
Default Converter
All enumeration classes based on built-in classes have a default conversion method to convert built-in data to an enumeration member of that class. For example:
a = Boolean(1) a =
12-20
Yes
An enumerated class also accepts enumeration members of its own class as input arguments:
Boolean(a) ans = Yes
Nonscalar inputs to the converter method return an object of the same size:
Boolean([0,1]) ans = No Yes
Create an empty enumeration array using the empty static method:

Boolean.empty ans = 0x0 empty Boolean enumeration.
12-21
12
Enumerations
Mutable (Handle) vs. Immutable (Value) Enumeration Members

In this section... Basic Knowledge on page 12-22 Selecting Handle- or Value-Based Enumerations on page 12-22 Value-Based Enumeration Classes on page 12-22 Handle-Based Enumeration Classes on page 12-24 Using Enumerations to Represent a State on page 12-28
Basic Knowledge
The material presented in this section builds on an understanding of the information provided in the following sections. Comparing Handle and Value Classes on page 5-2 enumeration function displays enumeration names See for general information about these two kinds of classes.
Selecting Handle- or Value-Based Enumerations

Use a handle enumeration when you want to enumerate a set of objects whose state might change over time. Use a value enumeration to enumerate a set of abstract (and immutable) values.
Value-Based Enumeration Classes

A value-based enumeration class has a fixed set of specific values. You cannot modify these values by changing the values of properties because doing so expands or changes the fixed set of values for this enumeration class.
12-22
Inherited Property SetAccess Must Be Immutable

Value-based enumeration class implicitly define the SetAccess attributes of all properties as immutable. You cannot set the SetAccess attribute to any other value. However, all superclass properties must explicitly define property SetAccess as immutable. See Property Attributes on page 6-8 for more information on property attributes.
Enumeration Members Remain Constant

When you create an instance of a value-based enumeration class, this instance is unique until the class is cleared and reloaded. For example, given the following class:
MATLAB considers a and b as equivalent:

a = WeekDays.Monday; b = WeekDays.Monday; isequal(a,b) ans = 1 a == b ans = 1
Enumeration Member Properties Remain Constant

Value-based enumeration classes that define properties are immutable. For example, the Colors enumeration class associates RGB values with color names.
12-23
12
Enumerations
classdef Colors properties R = 0; G = 0; B = 0; end methods function c = Colors(r, g, b) c.R = r; c.G = g; c.B = b; end end enumeration Red (1, 0, 0) Green (0, 1, 0) Blue (0, 0, 1) end end
The constructor assigns the input arguments to R, G, and B properties:

red = Colors.Red;
You cannot change a property value:

red.G = 1; Setting the 'G' property of the 'Colors' class is not allowed.
Handle-Based Enumeration Classes

Handle-based enumeration classes that define properties are mutable. Derive enumeration classes from the handle class when you must be able to change property values on instances of that class. Note You cannot derive an enumeration class from matlab.mixin.Copyable because the number of instances you can create are limited to the ones defined inside the enumeration block.
12-24
An Enumeration Member Remains Constant

Given a handle-based enumeration class with properties, changing the property value of an instance causes all references to that instance to reflect the changed value. For example, the HandleColors enumeration class associates RGB values with color names, the same as the Colors class in the previous example. However, HandleColors derives from handle:
classdef HandleColors < handle % Enumeration class derived from handle properties R = 0; G = 0; B = 0; end methods function c = HandleColors(r, g, b) c.R = r; c.G = g; c.B = b; end end enumeration Red (1, 0, 0) ... % Other colors omitted end end
Create an instance of HandleColors.Red and return the value of the R property:

a = HandleColors.Red; a.R ans = 1
12-25
12
Enumerations
MATLAB constructs the HandleColors.Red enumeration member, which sets the R property to 1, the G property to 0, and the B property to 0. Change the value of the R property to .8:
a.R = .8;
After setting the value of the R property to .8, create another instance, b, of HandleColors.Red:
b = HandleColors.Red; b.R ans = 0.8000
The value of the R property of the newly created instance is also 0.8000. The MATLAB session has only one value for any enumeration member at any given time. Clearing the workspace variables does not change the current definition of the enumeration member HandleColors.Red:
clear a = HandleColors.Red; a.R ans = 0.8000
Clear the class to reload the definition of the HandleColors class (see clear classes):
clear classes a = HandleColors.Red; a.R
12-26
ans = 1
If you do not want to allow reassignment of a given property value, set that propertys SetAccess attribute to immutable. See Property Attributes on page 6-8 for more information about property attributes.
Equality of Handle-Based Enumerations

Suppose you assign two variables to a particular enumeration member:
a = HandleColors.Red; b = HandleColors.Red;
You can compare a and b using isequal:

>> isequal(a,b) ans = 1
The property values of a and b are the same, so isequal returns true. However, unlike nonenumeration handle classes, a and b are the same handle because there is only one enumeration member. Determine handle equality using == (the handle eq method).
>> a == b ans = 1
See the handle eq method for information on how isequal and == differ when used with handles.
12-27
12
Enumerations
Using Enumerations to Represent a State

The MachineState class defines two enumeration members to represent the state of a machine, either running or not running.
classdef MachineState enumeration Running NotRunning end end
The Machine class represents a machine with start and stop operations. The MachineState enumerations are easy to work with because of their eq and char methods, and they result in code that is easy to read.
classdef Machine < handle properties (SetAccess = Private) State = MachineState.NotRunning; end methods function start(machine) if machine.State == MachineState.NotRunning machine.State = MachineState.Running; end disp (machine.State.char) end function stop(machine) if machine.State == MachineState.Running machine.State = MachineState.NotRunning; end disp (machine.State.char) end end end
Create a Machine object and call start and stop methods:

% Create a Machine object >> m = Machine; % Start the machine
12-28
>> m.start Running % Stop the machine >> m.stop NotRunning
12-29
12
Enumerations
Enumerations That Encapsulate Data

In this section... Basic Knowledge on page 12-30 Store Data in Properties on page 12-30
Basic Knowledge
The material presented in this section builds on an understanding of the information provided in the following sections. Fundamental MATLAB Classes for information on MATLAB built-in classes. enumeration function displays enumeration names
Store Data in Properties

Note Enumeration classes that subclass built-in numeric or logical classes cannot define or inherit properties. See Enumerations Derived from Built-In Types on page 12-16 for more information on this kind of enumeration class. Define properties in an enumeration class if you want to associate specific data with enumeration members, but do not need to inherit arithmetic, ordering, or other operations that MATLAB defines for specific built-in classes.
Representing Colors
Suppose you want to use a particular set of colors in all your graphs. You can define an enumeration class to represent the RGB values of the colors in your color set. The Colors class defines names for the colors, each of which uses the RGB values as arguments to the class constructor:
classdef Colors properties R = 0; G = 0; B = 0;
12-30
end methods function c = Colors(r, g, b) c.R = r; c.G = g; c.B = b; end end enumeration Blueish (18/255,104/255,179/255) Reddish (237/255,36/255,38/255) Greenish (155/255,190/255,61/255) Purplish (123/255,45/255,116/255) Yellowish (1,199/255,0) LightBlue (77/255,190/255,238/255) end end
Suppose you want to specify the new shade of red named Reddish:
a = Colors.Reddish; a.R ans = 0.9294 a.G ans = 0.1412 a.B ans = 0.1490
Use these values by accessing the enumeration members properties. For example, the myPlot function accepts a Colors enumeration member as an
12-31
12
Enumerations
input argument and accesses the RGB values defining the color from the property values.
function h = myPlot(x,y,LineColor) % Simple plotting function h = line('XData',x,'YData',y); r = LineColor.R; g = LineColor.G; b = LineColor.B; set(h,'Color',[r g b]) end
Create a plot using a reddish color line:

r = Colors.Reddish; h = myPlot(1:10,1:10,r);
The Colors class encapsulates the definitions of a standard set of colors. These definitions can change in the Colors class without affecting functions that use the Colors enumerations.
Enumerations Defining Categories

Suppose the Cars class defines categories used to inventory automobiles. The Cars class derives from the CarPainter class, which derives from handle. The abstract CarPainter class defines a paint method, which modifies the Color property if a car is painted another color. The Cars class uses Colors enumerations to specify a finite set of available colors. The exact definition of any given color can change independently of the Cars class.
classdef Cars < CarPainter enumeration Hybrid (2,'Manual',55,Colors.Reddish) Compact(4,'Manual',32,Colors.Greenish) MiniVan(6,'Automatic',24,Colors.Blueish) SUV (8,'Automatic',12,Colors.Yellowish) end properties (SetAccess = private) Cylinders
12-32
Transmission MPG Color end methods function obj = Cars(cyl,trans,mpg,colr) obj.Cylinders = cyl; obj.Transmission = trans; obj.MPG = mpg; obj.Color = colr; end function paint(obj,colorobj) if isa(colorobj,'Colors') obj.Color = colorobj; else [~,cls] = enumeration('Colors'); disp('Not an available color') disp(cls) end end end end
The CarPainter class requires its subclasses to define a method called paint:
classdef CarPainter < handle methods (Abstract) paint(carobj,colorobj) end end
Suppose you define an instance of the Cars class:

c1 = Cars.Compact;
The color of this car is Greenish, as defined by the Colors.Greenish enumeration:

c1.Color
12-33
12
Enumerations
ans = Greenish
Use the paint method to change the car color:

c1.paint(Colors.Reddish) c1.Color ans = Reddish
12-34
Saving and Loading Enumerations

In this section... Basic Knowledge on page 12-35 Built-In and Value-Based Enumeration Classes on page 12-35 Simple and Handle-Based Enumeration Classes on page 12-35 Causes: Loading as Struct Instead of Object on page 12-36
Basic Knowledge
See the save and load functions and Understanding the Save and Load Process on page 11-2 for general information on saving and loading objects. See the enumeration function to list enumeration names.
Built-In and Value-Based Enumeration Classes

When you save enumerations that derive from built-in classes or that are value-based classes with properties, MATLAB saves the names of the enumeration members and the definition of each member. When loading these types of enumerations, MATLAB preserves names over underlying values. If the saved named value is different from the current class definition, MATLAB uses the value defined in the current class, and then issues a warning.
Simple and Handle-Based Enumeration Classes

When you save simple enumerations (those having no properties, superclasses, or values associated with the member names) or those enumerations derived from the handle class, MATLAB saves the names and any underlying values. However, when loading these types of enumerations, MATLAB does not check the values associated with the names in the current class definition. This behavior results from the fact that simple enumerations have no underlying values and handle-based enumerations can legally have values that are different than those defined by the class.
12-35
12
Enumerations
Causes: Loading as Struct Instead of Object

The addition of a new named value or a new property made to a class subsequent to saving an enumeration does not trigger a warning during load. If there are changes to the enumeration class definition that do not prevent MATLAB from loading the object (that is, all of the named values in the MAT-File are present in the modified class definition), then MATLAB issues a warning that the class has changed and loads the enumeration. In the following cases, MATLAB issues a warning and loads as much of the saved data as possible as a struct: MATLAB cannot find the class definition The class is no longer an enumeration class MATLAB cannot initialize the class There is one or more enumeration member in the loaded enumeration that is not in the class definition For value-based enumerations with properties, a property exists in the file, but is not present in the class definition
Struct Fields
The returned struct has the following fields: ValueNames A cell array of strings, one per unique value in the enumeration array. Values An array of the same dimension as ValueNames containing the corresponding values of the enumeration members named in ValueNames. Depending on the kind of enumeration class, Values can be one of the following:
If the enumeration class derives from a built-in class, the array is of the built-in class and the values in the array are the underlying values of each enumeration member. Otherwise, a struct array representing the property name property values pairs of each enumeration member. For simple and handle-based enumerations, the struct array has no fields.
12-36
ValueIndices a uint32 array of the same size as the original enumeration. Each element is an index into the ValueNames and Values arrays. The content of ValueIndices represents the value of each object in the original enumeration array.
12-37
12
Enumerations
12-38
13
Constant Properties
13
Constant Properties
Properties with Constant Values

In this section... Defining Named Constants on page 13-2 Constant Property Assigned a Handle Object on page 13-4 Constant Property Assigned Any Class Instance on page 13-4
Defining Named Constants

Use constant properties to define constant values that you can access by name. Create a class with constant properties by declaring the Constant attribute in the property blocks. Setting the Constant attribute means that, once initialized to the value specified in the property block, the value cannot be changed.
Assigning Values to Constant Properties

Assign any value to a Constant property, including a MATLAB expression. For example:
classdef NamedConst properties (Constant) R = pi/180; D = 1/NamedConst.R; AccCode = '0145968740001110202NPQ'; RN = rand(5); end end
MATLAB evaluates the expressions when loading the class (when you first reference a constant property from that class). Therefore, the values MATLAB assigns to RN are the result of a single call to the rand function and do not change with subsequent references to NamedConst.RN. Calling clear classes causes MATLAB to reload the class and reinitialize the constant properties.
Referencing Constant Properties

Refer to the constant using the class name and the property name:
13-2
ClassName.PropName
For example, to use the NamedConst class defined in the previous section, reference the constant for the degree to radian conversion, R:
radi = 45*NamedConst.R radi = 0.7854
Constants In Packages
To create a library for constant values that you can access by name, first create a package folder, and then define the various classes to organize the constants you want to provide. For example, to implement a set of constants that are useful for making astronomical calculations, define a AstroConstants class in a package called constants:
+constants/AstroConstants/AstroConstants.m
The class defines a set of Constant properties with values assigned:

classdef AstroConstants properties (Constant) C = 2.99792458e8; G = 6.67259; Me = 5.976e24; Re = 6.378e6; end end
% % % %
m/s m/kgs Earth mass (kg) Earth radius (m)
To use this set of constants, reference them with a fully qualified class name. For example, the following function uses some of the constants defined in AstroContants:
function E = energyToOrbit(m,r) E = constants.AstroConstants.G * constants.AstroConstants.Me * m * ... (1/constants.AstroConstants.Re-0.5*r); end
13-3
13
Constant Properties
Importing the package into the function eliminates the need to repeat the package name (see import):
function E = energyToOrbit(m,r) import constants.*; E = AstroConstants.G * AstroConstants.Me * m * ... (1/AstroConstants.Re - 0.5 * r); end
Constant Property Assigned a Handle Object

If a class defines a constant property with a value that is a handle object, you can assign values to the handle objects properties. However, you must create a local variable to access the handle object. For example, the ConstMapClass class defines a constant property. The value of the constant property is a handle object (a containers.Map object).
classdef ConstMapClass < handle properties (Constant) ConstMapProp = containers.Map; end end
To assign the current date to the Date key, first return the handle from the constant property, and then make the assignment using the local variable on the left side of the assignment statement:
localMap = ConstantMapClass.ConstMapProp localMap('Date') = datestr(clock);
You cannot use a reference to a constant property on the left side of an assignment statement. For example, MATLAB interprets the following statement as the creation of a struct named ConstantMapClass with a field ConstMapProp:
ConstantMapClass.ConstMapProp('Date') = datestr(clock);
Constant Property Assigned Any Class Instance

You can assign an instance of the defining class to a constant property. MATLAB creates the instance assigned to the constant property when loading
13-4
the class. You can use this technique only when the defining class is a handle class. The MyProject is an example of such a class:
classdef MyProject < handle properties (Constant) ProjectInfo = MyProject; end properties Date Department ProjectNumber end methods (Access = private) function obj = MyProject obj.Date = datestr(clock); obj.Department = 'Engineering'; obj.ProjectNumber = 'P29.367'; end end end
Reference property data via the Constant property:

MyProject.ProjectInfo.Date ans = 18-Apr-2002 09:56:59
Because MyProject is a handle class, you can get the handle to the instance that is assigned to the constant property:
p = MyProject.ProjectInfo;
Access the data in the MyProject class using this handle:

p.Department ans =
13-5
13
Constant Properties
Engineering
Modify the nonconstant properties of the MyProject class using this handle:
p.Department = 'Quality Assurance'; p is a handle to the instance of MyProject that is assigned to the ProjectInfo
constant property:
MyProject.ProjectInfo.Department ans = Quality Assurance
Clearing the class results in the assignment of a new instance of MyProject to the ProjectInfo property.
clear MyProject MyProject.ProjectInfo.Department ans = Engineering
You can assign an instance of the defining class as the default value of a property only when the property is declared as Constant
13-6
14
Information from Class Metadata
Class Metadata on page 14-2 Inspecting Class and Object Metadata on page 14-5 Finding Objects with Specific Values on page 14-9 Getting Information About Properties on page 14-14 Find Default Values in Property Metadata on page 14-21
14
Class Metadata
In this section... What Is Class Metadata? on page 14-2 The meta Package on page 14-2 Metaclass Objects on page 14-3
What Is Class Metadata?

Class metadata is information about class definitions that is available from instances of metaclasses. Use metaclass objects to obtain information about class definitions without the need to create instances of the class itself. Each block in a class definition has an associated metaclass that defines the attributes for that block. Each attribute corresponds to a property in the metaclass. An instance of a metaclass has values assigned to each property that correspond to the values of the attributes of the associated class block. Metadata enables the programmatic inspection of classes. Tools such as property inspectors, debuggers, and so on, use these techniques.
The meta Package

The meta package contains metaclasses that MATLAB uses for the definition of classes and class components. The class name indicates the component described by the metaclass:
meta.package meta.class meta.property meta.DynamicProperty meta.EnumeratedValue meta.method meta.event
Each metaclass has properties, methods, and events that contain information about the class or class component. See meta.package, meta.class,
14-2
Class Metadata
meta.property, meta.DynamicProperty, meta.EnumeratedValue, meta.method and meta.event for more information on these metaclasses.
Metaclass Objects
Creating Metaclass Objects
You cannot instantiate metaclasses directly by calling the respective class constructor. Create metaclass objects from class instances or from the class name. ?ClassName Returns a meta.class object for the named class. Use meta.class.fromName with class names stored as characters in variables. meta.class.fromName('ClassName') returns the meta.class object for the named class (meta.class.fromName is a meta.class method). metaclass(obj) Returns a metaclass object for the class instance (metaclass)
% create metaclass object from class name using the ? operator mobj = ?classname; % create metaclass object from class name using the fromName method mobj = meta.class.fromName('classname'); % create metaclass object from class instance obj = myClass; mobj = metaclass(obj);
The metaclass function returns the meta.class object (that is, an object of the meta.class class). You can obtain other metaclass objects (meta.property, meta.method, and so on) from the meta.class object. Note Metaclass is a term used here to refer to all of the classes in the meta package. meta.class is a class in the meta package whose instances contain information about MATLAB classes. Metadata is information about classes contained in metaclasses.
14-3
14
Metaclass Object Lifecycle

When you change a class definition, MATLAB attempts to reload the class definition. If instances of the class exist, MATLAB cannot load the new class definition until you delete those instances. However, MATLAB does not allow existing metaclass objects to prevent the reloading of a new class definition. If you change a class definition while metaclass objects of that class exist, MATLAB deletes the metaclass objects and their handles become invalid. See Modifying and Reloading Classes on page 3-43 for information on how to modify and reload classes.
Using Metaclass Objects

Here are ways to access the information in metaclass objects: Obtain a meta.class object from a class definition (using ?) or from a class instance (using metaclass). Use the meta.class properties, methods, and events to obtain information about the class or class instance from which you obtained the meta.class object. For example, get other metaclass objects, such as the meta.properties objects defined for each of the class properties. See the following sections for examples that show how to use metadata: Inspecting Class and Object Metadata on page 14-5 Finding Objects with Specific Values on page 14-9 Getting Information About Properties on page 14-14 Find Default Values in Property Metadata on page 14-21
14-4
Inspecting Class and Object Metadata

In this section... Inspecting a Class on page 14-5 Metaclass EnumeratedValues Property on page 14-7
Inspecting a Class
The EmployeeData class is a handle class with two properties, one of which has private Access and defines a set access method.
classdef EmployeeData < handle properties EmployeeName end properties (Access = private) EmployeeNumber end methods function obj = EmployeeData(name,ss) if nargin > 0 obj.EmployeeName = name; obj.EmployeeNumber = ss; end end function set.EmployeeName(obj,name) if ischar(name) obj.EmployeeName = name; else error('Employee name must be a text string') end end end end
Inspecting the Class Definition

Using the EmployeeData class, create a meta.class object using the ? operator:
14-5
14
mc = ?EmployeeData;
Determine from what classes EmployeeData derives:

a = mc.SuperclassList; % a is an array of meta.class objects a.Name ans = handle
The EmployeeData class has only one superclass. For classes having more than one superclass, a would contain a meta.class object for each superclass. Use an indexed reference to refer to any particular superclass:
a(1).Name
or, directly from mc:

mc.SuperclassList(1).Name ans = handle
Inspecting Properties
Find the names of the properties defined by the EmployeeData class. First obtain an array of meta.properties objects from the meta.class PropertyList property.
mpArray = mc.PropertyList;
The length of mpArray indicates there are two meta.property objects, one for each property defined by the EmployeeData class:
length(mpArray) ans = 2
Now get a meta.property object from the array:
14-6
prop1 = mpArray(1); prop1.Name ans = EmployeeName
The Name property of the meta.property object identifies the class property represented by that meta.property object. Query other meta.property object properties to determine the attributes of the EmployeeName properties.
Inspecting an Instance of a Class

Create an EmployeeData object and determine property access settings:
EdObj = EmployeeData('My Name',1234567); mcEdObj = metaclass(EdObj); mpArray = mcEdObj.PropertyList; EdObj.(mpArray(1).Name) % Dynamic field names work with objects ans = My Name EdObj.(mpArray(2).Name) Getting the 'EmployeeNumber' property of the 'EmployeeData' class is not allowed. mpArray(2).GetAccess ans = private
Obtain a function handle to the property set access function:

mpArray(1).SetMethod ans = @D:\MyDir\@EmployeeData\EmployeeData.m>EmployeeData.set.EmployeeName
Metaclass EnumeratedValues Property

The meta.class EnumeratedValues property contains an array of meta.EnumeratedValue objects, one for each enumeration member. Use the meta.EnumeratedValue Name property to obtain the enumeration member names defined by an enumeration class. For example, given the WeekDays enumeration class:
14-7
14
Query enumeration names from the meta.class object:

mc = ?WeekDays; mc.EnumerationMemberList(2).Name ans = Tuesday
14-8
Finding Objects with Specific Values

In this section... Find Handle Objects on page 14-9 Find by Attribute Settings on page 14-10
Find Handle Objects

Use the handle class findobj method to find objects that have properties with specific values. For example, the following class defines a PhoneBook object to represent a telephone book entry in a data base. The PhoneBook class subclasses the dynamicprops class, which derives from handle.
classdef PhoneBook < dynamicprops properties Name Address Number end methods function obj = PhoneBook(n,a,p) obj.Name = n; obj.Address = a; obj.Number = p; end end end
Assume three of the PhoneBook entries in the database are:

PB(1) = PhoneBook('Nancy Vidal','123 Washington Street','5081234567'); PB(2) = PhoneBook('Nancy Vidal','123 Main Street','5081234568'); PB(3) = PhoneBook('Nancy Wong','123 South Street','5081234569');
One of these three PhoneBook objects has a dynamic property:

PB(2).addprop('HighSpeedInternet'); PB(2).HighSpeedInternet = '1M';
14-9
14
Find Property/Value Pairs

Find the object representing employee Nancy Wong and display the name and number by concatenating the strings:
NW = findobj(PB,'Name','Nancy Wong'); [NW.Name,' - ',NW.Number] ans = Nancy Wong - 5081234569
Find Objects with Specific Property Names

Search for objects with specific property names using the -property option:
H = findobj(PB,'-property','HighSpeedInternet'); H.HighSpeedInternet ans = 1M
The -property option enables you to omit the value of the property and search for objects using only the property name.
Using Logical Expressions

Search for specific combinations of property names and values:
H = findobj(PB,'Name','Nancy Vidal','-and','Address','123 Main Street'); H.Number ans = 5081234568
Find by Attribute Settings

All metaclasses derive from the handle class so you can use the handle findobj method to find class members that have specific attribute settings.
14-10
For example, find the abstract methods in a class definition by searching the meta.class MethodList for meta.method objects with their Abstract property set to true:
% Use class name in string form because class is abstract mc = meta.class.fromName('ClassName'); % Search list of meta.method objects for those % methods that have their Abstract property set to true absMethods = findobj(mc.MethodList,'Abstract',true); methodNames = {absMethods.Name};
The cell array, methodNames, contains the names of the abstract methods in the class.
Find Properties That Have Public Get Access

Find the names of all properties in the containers.Map class that have public GetAccess: Get the meta.class object Use findobj to search the array of meta.property objects
mc = ?containers.Map; % findobj returns an array of meta.property objects % use braces to convert the comman separated list to a cell array mpArray = findobj(mc.PropertyList,'GetAccess','public'); % create cell array of property names names = {mpArray.Name};
Display the names of all containers.Map properties that have public GetAccess:
celldisp(names) names{1} = Count names{2} =
14-11
14
KeyType names{3} = ValueType
Find Static Methods

Determine if any containers.Map class methods are static:
isempty(findobj([mc.MethodList(:)],'Static',true)) ans = 0 findobj returns an array of meta.method objects for the static methods. In this case, isempty returns false, indicating there are static methods defined
by this class. You can get the names of any static methods from the meta.method array:
staticMethodInfo = findobj([mc.MethodList(:)],'Static',true); staticMethodInfo(:).Name ans = empty
The name of the static method (there is only one in this case) is empty. Here is the information from the meta.method object for the empty method:
staticMethodInfo method with properties: Name: 'empty' Description: 'Returns an empty object array of the given size' DetailedDescription: ''
14-12
Access: Static: Abstract: Sealed: Hidden: InputNames: OutputNames: DefiningClass:
'public' 1 0 0 1 {'varargin'} {'E'} [1x1 meta.class]
14-13
14
Getting Information About Properties

In this section... The meta.property object on page 14-14 How to Find Properties with Specific Attributes on page 14-18
The meta.property object

The meta.property class is useful for determining the settings of property attributes. The writable properties of a meta.property object correspond to the attributes of the associated property. The values of the writable meta.property properties correspond to the attribute setting specified in the class definition. For example, create a default containers.Map object and use the handle findprop method to get the meta.property object for the Count property:
mp = findprop(containers.Map,'Count') mp = property with properties: Name: Description: DetailedDescription: GetAccess: SetAccess: Dependent: Constant: Abstract: Transient: Hidden: GetObservable: SetObservable: AbortSet: GetMethod: SetMethod: 'Count' 'Number of pairs in the collection' '' 'public' 'private' 1 0 0 1 0 0 0 0 [] []
14-14
DefiningClass: [1x1 meta.class]
The preceding meta.property display shows that the default Map object Count property has public GetAccess and private SetAccess, is Dependent, and Transient. See Table of Property Attributes on page 6-8 for a list of property attributes. If you are working with a class that is not a handle class, get the meta.property objects from the meta.class object. All metaclasses are subclasses of the handle class. Use the metaclass function if you have an instance or the ? operator with the class name:
mc = ?containers.Map mc = class with properties: Name: Description: DetailedDescription: Hidden: Sealed: ConstructOnLoad: HandleCompatible: InferiorClasses: ContainingPackage: PropertyList: MethodList: EventList: EnumerationMemberList: SuperclassList: 'containers.Map' 'MATLAB Map Container' 'MATLAB Map Container' 0 0 1 1 {0x1 cell} [1x1 meta.package] [4x1 meta.property] [35x1 meta.method] [1x1 meta.event] [0x1 meta.EnumeratedValue] [1x1 meta.class]
The meta.class object property named PropertyList contains an array of meta.property objects, one for each property defined by the containers.Map class. For example, the name of the property associated with the meta.property object in element 1 is:
mc.PropertyList(1).Name
14-15
14
ans = Count
The meta.class object contains a meta.property object for hidden properties too. Compare the result with the properties function, which returns only public properties:
properties('containers.Map') Properties for class containers.Map: Count KeyType ValueType
The serialization property is Hidden and has its GetAccess and SetAccess attributes set to private. Therefore, the properties function does not list it. However, you can get information about this property from its associated meta.property object (which is the fourth element in the array of meta.property objects in this case):
mc.PropertyList(4) ans = class with properties: Name: Description: DetailedDescription: GetAccess: SetAccess: Dependent: Constant: Abstract: Transient: Hidden: GetObservable: SetObservable: AbortSet: 'serialization' 'Serialization property.' '' 'private' 'private' 0 0 0 0 1 0 0 0
14-16
GetMethod: [] SetMethod: [] DefiningClass: [1x1 meta.class]
Indexing Metaclass Objects

Access other metaclass objects directly from the meta.class object properties. For example, the statement:
mc = ?containers.Map;
returns a meta.class object:

class(mc) ans = meta.class
Referencing the PropertyList meta.class property returns an array with one meta.property object for each property of the containers.Map class:
class(mc.PropertyList) ans = meta.property
Each array element is a single meta.property object:

mc.Properties(1) ans = [1x1 meta.property]
The Name property of the meta.property object contains a character string that is the name of the property:
class(mc.PropertyList(1).Name) ans =
14-17
14
char
Apply standard MATLAB indexing to access information in metaclass objects. For example, because the meta.class PropertyList property contains an array of meta.property objects, the following expression accesses the first meta.property object in this array and returns the first and last (C and t) letters of the string contained in the meta.property Name property.
mc.PropertyList(1).Name([1 end]) ans = Ct
How to Find Properties with Specific Attributes

This example implements a function that finds properties with specific attribute settings. For example, find objects that define constant properties (Constant attribute set to true) or determine what properties are read-only (GetAccess = public, SetAccess = private). The findAttrValue function returns a cell array of property names that set the specified attribute. This function accesses information from metaclasses using these techniques: If input argument, obj, is a string, use the meta.class.fromName static method to get the meta.class object. If input argument, obj, is an object, use the metaclass function to get the meta.class object. Every property has an associated meta.property object. Obtain these objects from the meta.class PropertyList property. Use the handle class findprop method to determine if the requested property attribute is a valid attribute name. All property attributes are properties of the meta.property object. The statement, findobj(mp,'PropertyName') determines whether the meta.property object, mp, has a property called PropertyName.
14-18
Reference meta.property object properties using dynamic field names. For example, if attrName = 'Constant', then MATLAB converts the expression mp.(attrName) to mp.Constant The optional third argument enables you to specify the value of attributes whose values are not logical true or false (such as GetAccess and SetAccess).
function cl_out = findAttrValue(obj,attrName,varargin) % Determine if first input is object or class name if ischar(obj) mc = meta.class.fromName(obj); elseif isobject(obj) mc = metaclass(obj); end % Initialize and preallocate ii = 0; numb_props = length(mc.PropertyList); cl_array = cell(1,numb_props); % For each property, check the value of the queried attribute for c = 1:numb_props % Get a meta.property object from the meta.class object mp = mc.PropertyList(c); % Determine if the specified attribute is valid on this object if isempty (findprop(mp,attrName)) error('Not a valid attribute name') end attrValue = mp.(attrName); % If the attribute is set or has the specified value, % save its name in cell array if attrValue if islogical(attrValue) || strcmp(varargin{1},attrValue) ii = ii + 1; cl_array(ii) = {mp.Name}; end
14-19
14
end end % Return used portion of array cl_out = cl_array(1:ii); end
Find Property Attributes

Suppose you have the following containers.Map object:
mapobj = containers.Map({'rose','bicycle'},{'flower','machine'});
Find properties with private SetAccess:

findAttrValue(mapobj,'SetAccess','private') ans = 'Count' 'KeyType' 'ValueType' 'serialization'
Find properties with public GetAccess:

findAttrValue(mapobj,'GetAccess','public') ans = 'Count' 'KeyType' 'ValueType'
14-20
Find Default Values in Property Metadata

In this section... meta.property Object on page 14-21 meta.property Data on page 14-21
meta.property Object
Class definitions can specify explicit default values for properties (see Defining Default Values on page 3-11). You can determine if a class defines explicit default values for a property and what the value of the default is from the propertys meta.property object.
meta.property Data
Obtain the default value of a property from the propertys associated meta.property object. The meta.class object for a class contains a meta.property object for every property defined by the class, including properties with private and protected access. For example:
mc = ?MException; % meta.class object for MException class mp = mc.PropertyList; % Array of meta.property objects mp(1) % meta.property object for 'type' property ans = class with properties: Name: Description: DetailedDescription: GetAccess: SetAccess: Dependent: Constant: Abstract: Transient: Hidden: GetObservable: 'type' 'Type of error reporting' '' 'private' 'private' 0 0 0 0 0 1
14-21
14
SetObservable: AbortSet: GetMethod: SetMethod: HasDefault: DefaultValue: DefiningClass:
1 0 [] [] 1 {} [1x1 meta.class]
Two meta.property object properties provide information on default values: HasDefault True if class specifies a default value for the property, false if it does not. DefaultValue Contains the default value, if the class defines a default value for the property. These properties provide a programmatic way to obtain property default values without reading class definition files. Use these meta.property object properties to obtain property default values for built-in classes and classes defined in MATLAB code.
Querying a Default Value

The procedure for querying a default value involves:
1 Getting the meta.property object for the property whose default value
you want to query.

2 Testing the logical value of the meta.property HasDefault property to
determine if the property defines a default value. MATLAB returns an error when you query the DefaultValue property if the class does not define a default value for the property.
3 Obtaining the default value from the meta.property DefaultValue
property if the HasDefault value is true. Use the ? operator, the metaclass function, or the meta.class.fromName static method (works with string variable) to obtain a meta.class object. The meta.class object PropertyList property contains an array of meta.property objects. Identify which property corresponds to which meta.property object using the meta.property Name property.
14-22
For example, this class defines properties with default values:

classdef MyDefs properties Material = 'acrylic'; InitialValue = 1.0; end end
Follow these steps to obtain the default value defined for the Material property. Include any error checking that is necessary for your application.
1 Get the meta.class object for the class:
mc = ?MyDefs;
2 Get an array of meta.property objects from the meta.class PropertyList
property:
mp = mc.PropertyList;
3 The length of the mp array equals the number of properties. You can use the
meta.property Name property to find the property of interest: for k = 1:length(mp) if (strcmp(mp(k).Name,'Material')
4 Before querying the default value of the Material property, test the
HasDefault meta.property to determine if MyClass defines a default
property for this property:

if mp(k).HasDefault dv = mp(k).DefaultValue; end end end
The DefaultValue property is read only. Changing the default value in the class definition changes the value of DefaultValue property. You can query the default value of a property regardless of its access settings.
14-23
14
Abstract and dynamic properties cannot define default values. Therefore, MATLAB returns an error if you attempt to query the default value of properties with these attributes. Always test the logical value of the meta.property HasDefault property before querying the DefaultValue property to avoid generating an error.
Default Values Defined as Expressions

Class definitions can define property default values as MATLAB expressions (see Expressions in Class Definitions on page 4-9 for more information). MATLAB evaluates these expressions the first time the default value is needed, such as the first time you create an instance of the class. Querying the meta.property DefaultValue property causes MATLAB to evaluate a default value expression, if it had not yet been evaluated. Therefore, querying a property default value can return an error or warning if errors or warnings occur when MATLAB evaluates the expression. See Property With Expression That Errors on page 14-25 for an example.
Property With No Explicit Default Value

MyClass does not explicitly define a default value for the Foo property: classdef MyFoo properties Foo end end
The meta.property instance for property Foo has a value of false for HasDefault. The class does not explicitly define a default value for Foo. Therefore, attempting to access the DefaultValue property causes an error:
mc = ?MyFoo; mp = mc.PropertyList(1); mp.HasDefault ans = 0 dv = mp.DefaultValue;
14-24
No default value has been defined for property Foo
Abstract Property
MyClass defines the Foo property as Abstract: classdef MyAbst properties (Abstract) Foo end end
The meta.property instance for property Foo has a value of false for its HasDefault property because you cannot define a default value for an Abstract property. Attempting to access DefaultValue causes an error:
mc = ?MyAbst; mp = mc.PropertyList(1); mp.HasDefault ans = 0 dv = mp.DefaultValue; Property Foo is abstract and therefore cannot have a default value.
Property With Expression That Errors

MyPropEr defines the Foo property default value as an expression that errors. classdef MyPropEr properties Foo = sin(pie/2); end end
The meta.property instance for property Foo has a value of true for its HasDefault property because Foo does have a default value determined by the evaluation of the expression:
sin(pie/2)
14-25
14
However, this expression returns an error (pie is a function that creates a pie graph, not the value pi).
mc = ?MyPropEr; mp = mc.PropertyList(1); mp.HasDefault ans = 1 dv = mp.DefaultValue; Error using pie Not enough input arguments.
Querying the default value causes the evaluation of the expression and returns the error.
Property With Explicitly Defined Default Value of Empty ([])

MyEmptyProp assigns a default of [] (empty double) to the Foo property: classdef MyEmptyProp properties Foo = []; end end
The meta.property instance for property Foo has a value of true for its HasDefault property. Accessing DefaultValue returns the value []:
mc = ?MyEmptyProp; mp = mc.PropertyList(1); mp.HasDefault ans = 1 dv = mp.DefaultValue; dv = []
14-26
14-27
14
14-28
15
Specializing Object Behavior
Methods That Modify Default Behavior on page 15-2 Redefining Concatenation for Your Class on page 15-6 Converting Objects to Another Class on page 15-7 Indexed Reference and Assignment on page 15-9 Implementing Operators for Your Class on page 15-31
15
Methods That Modify Default Behavior

In this section... How to Customizing Class Behavior on page 15-2 Which Methods Control Which Behaviors on page 15-2 Overloading and Overriding Functions and Methods on page 15-4 Caution When Customizing Classes on page 15-5
How to Customizing Class Behavior

There are functions that MATLAB calls implicitly when you perform certain actions with objects. For example, a statement like [B(1);A(3)] involves indexed reference and vertical concatenation. You can change how user-defined objects behave by defining methods that control specific behaviors. To change a behavior, implement the appropriate method with the name and signature of the MATLAB function.
Which Methods Control Which Behaviors

The following table lists the methods to implement for your class and describes the behaviors that they control. Class Method to Implement Concatenating Objects
cat, horzcat, and vertcat
Description Customize behavior when concatenation objects See Subclasses of Built-In Types with Properties on page 10-59
Creating Empty Arrays

empty
Create empty arrays of the specified class. See Creating Empty Arrays on page 8-6
Displaying Objects
15-2
Class Method to Implement

disp display
Description Called when you enter disp(obj) on the command line Called when statements are not terminated by semicolons. disp is often used to implement display methods. See Overloading the disp Function on page 16-45
Converting Objects to Other Classes converters like double and

char
Convert an object to a MATLAB built-in class See The DocPolynom to Character Converter on page 17-7 and The DocPolynom to Double Converter on page 17-6
Indexing Objects
subsref and subsasgn
Enables you to create nonstandard indexed reference and indexed assignment See Indexed Reference and Assignment on page 15-9
end
Supports end syntax in indexing expressions using an object; e.g.,

A(1:end)
See Defining end Indexing for an Object on page 15-27

numel
Determine the number of elements in an array See Interactions with numel and subsref and subsasgn on page 15-5
size
Determine the dimensions in an array
15-3
15
Class Method to Implement

subsindex
Description Support using an object in indexing expressions See Using Objects as Indices on page 15-28

loadobj and saveobj
Customize behavior when loading and saving objects See Control Save and Load
Overloading and Overriding Functions and Methods

Overloading and overriding are terms that describe techniques for customizing class behavior. Here is how we use these terms in MATLAB.
Overloading
Overloading means that there is more than one function or method having the same name within the same scope. MATLAB dispatches to a particular function or method based on the dominant argument. For example, the timeseries class overloads the MATLAB plot function. When you call plot with a timeseries object as an input argument, MATLAB calls the timeseries class method named plot.
Overriding
Overriding means redefining a method inherited from a superclass. MATLAB dispatches to the most specific version of the method. That is, if the dominant argument is an instance of the subclass, then MATLAB calls the subclass method. Use the InferiorClasses attribute to control class precedence. See Class Attributes on page 4-6 for more information.
15-4
Caution When Customizing Classes

Many MATLAB functions depend on the behavior of other functions, like size and numel. Therefore, you must be careful to ensure that what is returned by an overloaded version of these functions is a correct and accurate representation of the size of an object array.
Interactions with numel and subsref and subsasgn

You might need to define a numel method to compensate when your class defines a specialized version of size.
subsref uses the value returned by numel to compute the number of expected output arguments returned by subsref from subscripted reference (i.e., nargout).
Similarly, subsasgn uses numel to compute the expected number of input arguments to be assigned using subsasgn (i.e., nargin). MATLAB determines the value of nargin for an overloaded subsasgn function from the value returned by numel, plus two (one for the variable to which you are making an assignment and one for the struct array of subscripts). If MATLAB produces errors when calling your classs overloaded subsref or subsagn methods because nargout is wrong for subsref or nargin is wrong for subsasgn, then you need to overload numel to return a value that is consistent with your implementation of these indexing functions. See Understanding size and numel on page 10-65 and Indexed Reference and Assignment on page 15-9 for more information on implementing subsref and subsagn methods.
15-5
15
Redefining Concatenation for Your Class

Default Concatenation
You can concatenate objects into arrays. For example, suppose you have three instances of the class MyClass, obj1, obj2, obj3. You can form various arrays with these objects using brackets. Horizontal concatenation calls horzcat:
HorArray = [obj1,obj2,obj3]; HorArray is a 1by3 array of class MyClass. You can concatenate the objects along the vertical dimension, which calls vertcat: VertArray = [obj1;obj2;obj3] VertArray is a 3by1 array of class MyClass. You can use the cat function to concatenate arrays along different dimensions. For example: ndArray = cat(3,HorArray,HorArray); ndArray is a 1by3by2 array.
Overload horzcat, vertcat, and cat to produce specialized behaviors in your class. You must overload both horzcat and vertcat whenever you want to modify object concatenation because MATLAB uses both functions for any concatenation operation.
Example of horzcat and vertcat

Subclasses of Built-In Types with Properties on page 10-59
15-6
Converting Objects to Another Class
Converting Objects to Another Class

Why Implement a Converter
You can convert an object of one class to an object of another class. A converter method has the same name as the class it converts to, such as char or double. Think of a converter method as an overloaded constructor method of another classit takes an instance of its own class and returns an object of a different class. Converters enable you to: Use methods defined for another class Ensure that expressions involving objects of mixed class types execute properly Control how instances are interpreted in other contexts Suppose you define a polynomial class. If you create a double method for the polynomial class, you can use it to call other functions that require inputs of type double.
p = polynomial(...); dp = double(p); roots(dp) p is a polynomial object, double is a method of the polynomial class, and roots is a standard MATLAB function whose input arguments are the
coefficients of a polynomial.
Converters and Subscripted Assignment

When you make a subscripted assignment statement such as:
A(1) = myobj;
MATLAB compares the class of the Right-Hand-Side (RHS) variable to the class of the Left-Hand-Side (LHS) variable. If the classes are different, MATLAB attempts to convert the RHS variable to the class of the LHS variable. To do this, MATLAB first searches for a method of the RHS class
15-7
15
that has the same name as the LHS class. Such a method is a converter method, which is similar to a typecast operation in other languages. If the RHS class does not define a method to convert from the RHS class to the LHS class, then MATLAB software calls the LHS class constructor and passes it to the RHS variable. For example, suppose you make the following assignments:
A(1) = objA; % Object of class ClassA A(2) = objB; % Object of class ClassB
MATLAB attempts to call a method of ClassB named ClassA. If no such converter method exists, MATLAB software calls the ClassA constructor, passing objB as an argument. If the ClassA constructor cannot accept objB as an argument, then MATLAB returns an error. You can create arrays of objects of different classes using cell arrays (see cell for more information on cell arrays).
Examples of Converter Methods

See the following sections for examples of converter methods: The DocPolynom to Double Converter on page 17-6 The DocPolynom to Character Converter on page 17-7 Subclasses of Built-In Types with Properties on page 10-59
15-8
Indexed Reference and Assignment

In this section... Overview on page 15-9 Default Indexed Reference and Assignment on page 15-9 What You Can Modify on page 15-11 subsref and subsasgn Within Class Methods Built-In Called on page 15-12 Understanding Indexed Reference on page 15-14 Avoid Overriding Access Attributes on page 15-17 Understanding Indexed Assignment on page 15-19 A Class with Modified Indexing on page 15-22 Defining end Indexing for an Object on page 15-27 Using Objects as Indices on page 15-28
Overview
This section describes how indexed reference and assignment work in MATLAB, and provides information on the behaviors you can modify. There are also examples of classes that modify the default indexing behavior. MATLAB provides support for object array indexing by default and many class designs will require no modification to this behavior. The information in this section can help you determine if modifying object indexing is useful for your class design and can show you how to approach those modifications.
Default Indexed Reference and Assignment

MATLAB arrays enable you to reference and assign elements of the array using a subscripted notation that specifies the indices of specific array elements. For example, suppose you create two arrays of numbers (using randi and concatenation).
% Create a 3-by-4 array of integers between 1 and 9 A = randi(9,3,4)
15-9
15
A = 4 8 5 7 4 2 6 3 7 5 7 7 % Create a 1-by-3 array of the numbers 3, 6, 9 B = [3 6 9];
You can reference and assign elements of either array using index values in parentheses:
B(2) = A(3,4); B B = 3 7
When you execute a statement that involves indexed reference:

C = A(3,4);
MATLAB calls the built-in subsref function to determine how to interpret the statement. Similarly, if you execute a statement that involves indexed assignment:
C(4) = 7;
MATLAB calls the built-in subsasgn function to determine how to interpret the statement. The MATLAB default subsref and subsasgn functions also work with user-defined objects. For example, suppose you want to create an array of objects of the same class:
for k=1:3 objArray(k) = MyClass; end
Referencing the second element in the object array, objArray, returns the object constructed when k = 2:
D = objArray(2);
15-10
class(D) ans = MyClass
You also can assign an object to an array of objects of the same class, or an uninitialized variable (see Creating Empty Arrays on page 8-6 for related information):
newArray(3,4) = D;
Arrays of objects behave much like numeric arrays in MATLAB. You do not need to implement any special methods to provide this behavior with your class. For general information about array indexing, see Matrix Indexing.
What You Can Modify

You can modify your classs default indexed reference and/or assignment behavior by implementing class methods called subsref and subsasgn. For syntax description, see their respective reference pages. Keep in mind that once you add a subsref or subsasgn method to your class, then MATLAB calls only the class method, not the built-in function. Therefore, you must implement in your class method all of the indexed reference and assignment operations that you want your class to support. This includes: Dot notation calls to class methods Dot notation reference and assignment involving properties Any indexing using parentheses '()' Any indexing using braces '{}' While implementing subsref and subsasgn methods gives you complete control over the interpretation of indexing expressions for objects of your class, it can be complicated to provide the same behavior that MATLAB provides by default.
15-11
15
When to Modify Indexing Behavior

The default indexing supported by MATLAB for object arrays and dot notation for access to properties and methods enables user-defined objects to behave like intrinsic classes, such as double and struct. For example, suppose you define a class with a property called Data that contains an array of numeric data. A statement like:
obj.Data(2,3)
returns the value contained in the second row, third column of the array. If you have an array of objects, you can use an expression like:
objArray(3).Data(4:end)
to return the fourth through last elements in the array contained in the Data property of the third object in the object array, objArray. Modify the default indexing behavior when your class design requires behavior that is different from that provided by MATLAB by default.
subsref and subsasgn Within Class Methods Built-In Called

MATLAB does not call class-defined subsref or subsasgn methods for indexed reference and assignment within the classs own methods. Within class methods, MATLAB always calls the built-in subsref and subsasgn functions regardless of whether the class defines its own methods. This is true within the class-defined subsref and subsasgn methods as well. For example, within a class method, this dot reference:
% Calls built-in subsref obj.Prop
calls the built-in subsref function. To call the class-defined subsref method, use:
% Calls overloaded subsref subsref(obj,substruct('.','Prop'))
15-12
Whenever a class method requires the functionality of the class-defined subsref or subsasgn method, it must call the overloaded methods with function calls rather than using the operators, '()', '{}', or '.'. For example, suppose you define a polynomial class with a subsref method that causes the polynomial to be evaluated with the value of the independent variable equal to the subscript. This statement defines the polynomial with its coefficients:
p = polynom([1 0 -2 -5]);
The MATLAB expression for the resulting polynomial is:

x^3 - 2*x - 5
The following subscripted expression returns the value of the polynomial at x = 3:

p(3) ans = 16
Suppose that you want to use this feature in another class method. To do so, call the subsref function directly. The evalEqual method accepts two polynom objects and a value at which to evaluate the polynomials:
methods function ToF = evalEqual(p1,p2,x) % Create arguments for subsref subs.type = '()'; subs.subs = {x}; % Need to call subsref explicity y1 = subsref(p1,subs); y2 = subsref(p2,subs); if y1 == y2 ToF = true; else ToF = false; end end end
15-13
15
This behavior enables you to use standard MATLAB indexing to implement specialized behaviors. See A Class with Modified Indexing on page 15-22 for examples of how to use both built-in and class-modified indexing.
Understanding Indexed Reference

Object indexed references are in three forms parentheses, braces, and name:
A(I) A{I} A.name
Each of these statements causes a call by MATLAB to the subsref method of the class of A, or a call to the built-in subsref function, if the class of A does not implement a subsref method. MATLAB passes two arguments to subsref:
B = subsref(A,S)
The first argument is the object being referenced, A. The second argument, S, is a struct array with two fields: S.type is a string containing '()', {}', or '.' specifying the indexing type used. S.subs is a cell array or string containing the actual index or name. A colon used as an index is passed in the cell array as the string ':'. Ranges specified using a colon (e.g., 2:5) are expanded to 2 3 4 5. For example, the expression
A(1:4,:)
causes MATLAB to call subsref(A,S), where S is a 1-by-1 structure with

S.type = '()' S.subs = {1:4,':'} % A 2-element cell array % containing the numbers 1 2 3 4 and ":"
15-14
Returning the contents of each cell of S.subs gives the index values for the first dimension and a string ':' for the second dimension:
S.subs{:} ans = 1 ans = : 2 3 4
The default subsref returns all array elements in rows 1 through 4 and all of the columns in the array. Similarly, the expression
A{1:4}
uses
S.type ='{}' S.subs = {1:4} % A cell array % containing the numbers 1 2 3 4
The default subsref returns the contents of all cell array elements in rows 1 through 4 and all of the columns in the array. The expression
A.Name
calls subsref(A,S) where

S.type = '.' S.subs = 'Name' % The string 'Name'
The default subsref returns the contents of the Name field in the struct array or the value of the property Name if A is an object with the specified property name.
15-15
15
Complex Indexed References

These simple calls are combined for more complicated indexing expressions. In such cases, length(S) is the number of indexing levels. For example,
A(1,2).PropertyName(1:4)
calls subsref(A,S), where S is a 3-by-1 structure array with the values:

S(1).type = '()' S(1).subs = {1,2} S(2).type = '.' S(2).subs = 'PropertyName' S(3).type = '()' S(3).subs = {1:4}
Writing subsref
Your classs subsref method must interpret the indexing expressions passed in by MATLAB. Any behavior you want your class to support must be implemented by your subsref. However, your method can call the built-in subsref to handle indexing types that you do not want to change. You can use a switch statement to determine the type of indexing used and to obtain the actual indices. The following three code fragments illustrate how to interpret the input arguments. In each case, the function must return the value (B) that is returned by your subsref function. For a parentheses index:
% Handle A(n) switch S.type case '()' B = A(S.subs{:}); end
For a brace index:

% Handle A{n} switch S.type case '{}' % Determine what this indexing means to your class % E.g., CellProperty contained a cell array B = A.CellProperty{S.subs{:}}; end
15-16
While braces are used for cell arrays in MATLAB, your subsref method is free to define its own meaning for this syntax. For a name index, you might access property values. Method calls require a second level of indexing if there are arguments. The name can be an arbitrary string for which you take an arbitrary action:
switch S.type case '.' switch S.subs case 'name1' B = A.name1; case 'name2' B = A.name2; end end
Examples of subsref
These links show examples of classes that implement subsref methods: A Class with Modified Indexing on page 15-22 Subclasses of Built-In Types with Properties on page 10-59 A Class to Represent Hardware on page 10-70 The DocPolynom subsref Method on page 17-10 See also, Understanding size and numel on page 10-65
Avoid Overriding Access Attributes

Because subsref is a class method, it has access to private class members. Avoid inadvertently giving access to private methods and properties as you handle various types of reference. Consider this subsref method defined for a class having private properties, x and y:
classdef MyPlot
15-17
15
properties (Access = private) x y end properties Maximum Minimum Average end methods function obj = MyPlot(x,y) obj.x = x; obj.y = y; obj.Maximum = max(y); obj.Minimum = min(y); obj.Average = mean(y); end function B = subsref(A,S) switch S(1).type case '.' switch S(1).subs case 'plot' % Reference to A.x and A.y call built-in subsref B = plot(A.x,A.y); otherwise % Enable dot notation for all properties and methods B = A.(S.subs); end end end end end
This subsref enables users to use dot notation to perform an action (create a plot) using the name 'plot'. The statement:
obj = MyPlot(1:10,1:10); h = obj.plot;
calls the plot function and returns the handle to the graphics object.
15-18
You do not need to explicitly code each method and property name because the otherwise code in the inner switch block handles any name reference that you do not explicitly specify in case statements. However, using this technique exposes any private and protected class members via dot notation. For example, you can reference the private property, x, with this statement:
obj.x ans = 1 2 3 4 5 6 7 8 9 10
The same issue applies to writing a subsasgn method that enables assignment to private or protected properties. Your subsref and subsasgn methods might need to code each specific property and method name explicitly to avoid violating the class design.
Understanding Indexed Assignment

Object indexed assignments are in three forms parentheses, braces, and name:
A(I) = B A{I} = B A.name = B
Each of these statements causes a call by MATLAB to the subsasgn method of the class of A, or a call to the built-in function, if the class of A does not implement a subsasgn method. MATLAB passes three arguments to subsasgn:
A = subsasgn(A,S,B)
The first argument, A, is the object being assigned the value in the third argument B. The second argument, S, is a struct array with two fields:
15-19
15
S.type is a string containing '()', '{}', or '.' specifying the indexing type used. S.subs is a cell array or string containing the actual index or name. A colon used as an index is passed in the cell array as the string ':'. Ranges specified using a colon (e.g., 2:5) are expanded to 2 3 4 5. For example, the assignment statement:
A(2,3) = B;
generates a call to subsasgn: A = subsasgn(A,S,B) where S is:

S.type = '()' S.subs = {2,3}
The default subsasgn: Determines the class of A. If B is not the same class a A, then MATLAB tries to construct an object of the same class as A using B as an input argument (e.g., by calling a converter method, if one exists). If this attempt fails, MATLAB returns an error. If A and B are, or can be made, into the same class, then MATLAB assigns the value of B to the array element at row 2, column 3. If A does not exist before you execute the assignment statement, then MATLAB initializes the five array elements that come before A(2,3) with a default object of the class of A and B. For example, empty elements are initialized to zero in the case of a numeric array or an empty cell ([]) in the case of cell arrays. See Creating Empty Arrays on page 8-6 for more information on how MATLAB initializes empty arrays. Similarly, the expression
A{2,3} = B
uses
S.type ='{}' S.subs = {2,3} % A 2-element cell array containing the numbers 2 and 3
The default subsasgn:
15-20
Assigns B to the cell array element at row 2, column 3. If A does not exist before you execute the assignment statement, MATLAB initializes the five cells that come before A(2,3) with []. The result is a 2by3 cell array. The expression
A.Name = B
calls A = subsasgn(A,S,B) where

The default subsasgn: Assigns B to the struct field Name. If A does not exist before you execute the assignment statement, MATLAB creates a new struct variable, A with field Name and assigns the value of B to this field location. If struct A exists, but has no field Name, then MATLAB adds the field Name and assigns the value of B to the new field location. If struct A exists and has a Name field, then MATLAB assigns the value of B to Name. You can redefine all or some of these assignment behaviors by implementing a subsasgn method for your class.
Indexed Assignment to Objects

If A is an object, the expression:
A.Name = B
calls A = subsasgn(A,S,B) where

15-21
15
The default subsasgn: Attempts to assign B to the Name property. If the class of A does not have a Name property, MATLAB returns an error. If the Name property has restricted access (private or protected), MATLAB determines if the assignment is allowed based on the context in which the assignment is made. If the class of A defines a set method for property Name, MATLAB calls the set method. MATLAB applies all other property attributes before determining whether to assigning B to the property Name.
Complex Indexed Assignments

These simple calls are combined for more complicated indexing expressions. In such cases, length(S) is the number of indexing levels. For example,
A(1,2).PropertyName(1:4) = B
calls subsasgn(A,S,B), where S is a 3-by-1 structure array with the values:

S(1).type = '()' S(1).subs = {1,2} S(2).type = '.' S(2).subs = 'PropertyName' S(3).type = '()' S(3).subs = {1:4}
For an example of subsasgn method, see Specialized Subscripted Assignment subsasgn on page 15-25.
A Class with Modified Indexing

This example defines a class that modifies the default indexing behavior. It uses a combination of default indexing and specialized indexing. The example shows some useful techniques for implement subsref and subsasgn methods, but does not implement a fully robust class. You cannot, for example, concatenate objects into an array without adding other methods, such as horzcat, vertcat, cat, size, and perhaps other methods. See Subclasses of Built-In Types with Properties on page 10-59 for another example of a class that modifies indexing and concatenation behavior.
15-22
Class Description
The class has three properties: Data numeric test data Description description of test data Date date test was conducted Assume you have the following data (randi):
d = randi(9,3,4) d = 8 9 2 9 6 1 3 5 9 9 2 9
Create an instance of the class:

obj = MyDataClass(d,'Test001');
The constructor arguments pass the values for the Data and Description properties. The clock function assigns the value to the Date property from within the constructor. This approach captures the time and date information when the instance is created. Here is the basic code listing without the subsref and subsasgn methods.
classdef MyDataClass properties Data Description end properties (SetAccess = private) Date end methods function obj = MyDataClass(data,desc) if nargin > 0 obj.Data = data; end
15-23
15
if nargin > 1 obj.Description = desc; end obj.Date = clock; end end end
Specialized Subscripted Reference subsref

Use the default indexed reference behavior for scalar objects, and add the ability to index into the Data property with an expression like:
obj(2,3) ans = 5
This statement is the equivalent of:

obj.Data(2,3)
which the class also supports. Redefining '()' indexing as described here means you cannot create arrays of MyDataClass objects and use '()' indexing to access individual objects. Create only scalar objects. To achieve the design goals, the subsref method calls the builtin subsref for indexing of type '.' and defines its own version of '()' type indexing.
function sref = subsref(obj,s) % obj(i) is equivalent to obj.Data(i) switch s(1).type % Use the built-in subsref for dot notation case '.' sref = builtin('subsref',obj,s); case '()' if length(s)<2 % Note that obj.Data is passed to subsref
15-24
sref = builtin('subsref',obj.Data,s); return else sref = builtin('subsref',obj,s); end % No support for indexing using '{}' case '{}' error('MYDataClass:subsref',... 'Not a supported subscripted reference') end end
Specialized Subscripted Assignment subsasgn

The class supports the equivalent behavior in indexed assignment. You can assign values to the Data property by referencing only the object.
obj(2,3) = 9;
is equivalent to:
obj.Data(2,3) = 9;
Like the subsref method, the subsasgn method calls the builtin subsasgn for indexing of type '.' and defines its own version of '()' type indexing. Another useful approach is the use of the substruct function to redefine the index type and index subscripts struct that MATLAB passes to subsref and subsasgn.
function obj = subsasgn(obj,s,val) if isempty(s) && strcmp(class(val),'MYDataClass') obj = MyDataClass(val.Data,val.Description); end switch s(1).type % Use the built-in subsasagn for dot notation case '.' obj = builtin('subsasgn',obj,s,val); case '()' if length(s)<2
15-25
15
if strcmp(class(val),'MYDataClass') error('MYDataClass:subsasgn',... 'Object must be scalar') elseif strcmp(class(val),'double') % Redefine the struct s to make the call: obj.Data(i) snew = substruct('.','Data','()',s(1).subs(:)); obj = subsasgn(obj,snew,val); end end % No support for indexing using '{}' case '{}' error('MYDataClass:subsasgn',... 'Not a supported subscripted assignment') end end
Implementing Addition for Object Data plus

Allow direct addition of the Data property data by implementing a plus method:
function a = double(obj) a = obj.Data; end function c = plus(obj,b) c = double(obj) + double(b); end
For example, add a scalar to the object Data array:

% List Current value of Data obj(:,:) ans = 8 9 2 9 6 1 3 5 9 9 2 9
% Add 7 to the array
15-26
obj + 7 ans = 15 16 9 16 13 8 10 12 16 16 9 16
The MyDataClass double method provides a way to convert an object to an array of doubles. It is possible to add a MyDataClass object to another class of object, providing the other class implements a double method that also returns an array of doubles. MATLAB applies the rules of addition and returns errors for dimension mismatch, and so on.
Defining end Indexing for an Object

When you use end in an object indexing expression, such as A(4:end), the end function returns the index value corresponding to the last element in that dimension. Classes can overload the end function as a class method to implement specialized behavior. If your class defines an end method, MATLAB calls that method to determine how to interpret the expression. The end has the calling syntax:
ind = end(A,k,n)
where A is the object k is the index in the expression using the end syntax n is the total number of indices in the expression ind is the index value to use in the expression For example, consider the expression
A(end-1,:)
15-27
15
MATLAB calls the end method defined for the object A using the arguments
ind = end(A,1,2)
These arguments mean the end statement occurs in the first index element and there are two index elements. The end class method returns the index value for the last element of the first dimension (from which 1 is subtracted in this case). If your class implements an end method, ensure that it returns a value appropriate for the class.
The end Method for the MyDataClass Example

The end method for the MyDataClass example (see A Class with Modified Indexing on page 15-22) operates on the contents of the Data property. The objective of this method is to return a value that can replace end in any indexing expression, such as:
obj(4:end) obj.Data(2,3:end)
and so on. The following end function determines a positive integer value for end and returns it so that MATLAB can plug it into the indexing expression.
function ind = end(obj,k,n) szd = size(obj.Data); if k < n ind = szd(k); else ind = prod(szd(k:end)); end end
Using Objects as Indices

MATLAB can use objects as indices in indexed expressions. The rules of array indexing apply indices must be positive integers. Therefore, MATLAB must be able to derive a value from the object that is a positive integer, which it uses in the indexed expression.
15-28
Indexing expressions like X(A), where A is an object, cause MATLAB to call the default subsindex function, unless such an expression results in a call to an overloaded subsref or subsasgn method defined by the class of X. See Scenarios for Implementing Objects as Indices on page 15-29.
subsindex must return the value of the object as a zero-based integer index values in the range 0 to prod(size(X))-1).
Scenarios for Implementing Objects as Indices

If you want to enable indexing of one object by another object, such as X(A), you can implement this behavior in various ways: Define a subsindex method in the class of A, which converts A to an integer. MATLAB calls As subsindex method to perform default indexing operations (when the class of X does not overload the default subsref or subsasgn method). If the class of X overloads subsref or subsasgn, these methods can explicitly call the subsindex method of A. In this case, ensure that A implements a subsindex method with appropriate error checking in your program. If the class of X overloads subsref or subsasgn, these methods can contain code that determines an integer index value without relying on the class of A to implement a subsindex method.
Implementing subsindex
MATLAB calls the subsindex method defined for the object used as the index. For example, suppose you want to use object A to index into object B. B can be a single object or an array, depending on your objectives.
C = B(A);
A subsindex method implemented by class A might do something as simple as convert the object to double format to be used as an index, as shown in this sample code.
function ind = subsindex(obj) % Convert the object a to double format to be used % as an index in an indexing expression ind = double(obj);
15-29
15
end
Or, your class might implement a special converter method that returns a numeric value representing an object based on particular values of object properties.
function ind = subsindex(obj) % Return the value of an object property ind = obj.ElementPosition; end subsindex values are 0-based, not 1-based.
15-30
Implementing Operators for Your Class

In this section... Defining Operators on page 15-31 MATLAB Operators and Associated Functions on page 15-32
Defining Operators
You can implement MATLAB operators (+, *, >, etc.) to work with objects of your class. Do this by defining the associated class methods. Each operator has an associated function (e.g., the + operator has an associated plus.m function). You can implement any operator by creating a class method with the appropriate name. This method can perform whatever steps are appropriate for the operation being implemented. See MATLAB Operators and Associated Functions on page 15-32 for a complete list of operator functions.
Object Precedence in Operations

User-defined classes have a higher precedence than built-in classes. For example, if q is an object of class double and p is a user-defined class, MyClass, both of these expressions:
q + p p + q
generate a call to the plus method in the MyClass, if it exists. Whether this method can add objects of class double and class MyClass depends on how you implement it. When p and q are objects of different classes, MATLAB applies the rules of precedence to determine which method to use. Object Precedence in Expressions Using Operators on page 7-29 provides information on how MATLAB determines which method to call.
15-31
15
Examples of Operator Implementations

Defining Arithmetic Operators for DocPolynom on page 17-13 provides examples of overloaded operators.
MATLAB Operators and Associated Functions

The following table lists the function names for MATLAB operators. Implementing operators to work with arrays, (scalar expansion, vectorized arithmetic operations, an so on), might also require modifying indexing and concatenation. Operation
a + b a - b -a +a a.*b a*b a./b a.\b a/b a\b a.^b a^b a < b a > b a <= b a >= b
Method to Define
plus(a,b) minus(a,b) uminus(a) uplus(a) times(a,b) mtimes(a,b) rdivide(a,b) ldivide(a,b) mrdivide(a,b) mldivide(a,b) power(a,b) mpower(a,b) lt(a,b) gt(a,b) le(a,b) ge(a,b)
Description Binary addition Binary subtraction Unary minus Unary plus Element-wise multiplication Matrix multiplication Right element-wise division Left element-wise division Matrix right division Matrix left division Element-wise power Matrix power Less than Greater than Less than or equal to Greater than or equal to
15-32
Operation
a ~= b a == b a & b a | b ~a a:d:b a:b a' a.' command window output [a b] [a; b] a(s1,s2,...sn) a(s1,...,sn) = b b(a)
Method to Define
ne(a,b) eq(a,b) and(a,b) or(a,b) not(a) colon(a,d,b) colon(a,b) ctranspose(a) transpose(a) display(a) horzcat(a,b,...) vertcat(a,b,...) subsref(a,s) subsasgn(a,s,b) subsindex(a)
Description Not equal to Equality Logical AND Logical OR Logical NOT Colon operator Complex conjugate transpose Matrix transpose Display method Horizontal concatenation Vertical concatenation Subscripted reference Subscripted assignment Subscript index
15-33
15
15-34
16
Customizing Object Display
Custom Display Interface on page 16-2 How CustomDisplay Works on page 16-8 Role of the size Function In Custom Displays on page 16-11 Custom Display for Heterogeneous Arrays on page 16-13 Class with Default Object Display on page 16-15 How to Approach Display Customization on page 16-20 Customize Property Display on page 16-24 Customize Header, Property List, and Footer on page 16-28 Customize Display of Scalar Objects on page 16-35 Custom Display of Object Arrays on page 16-40 Overloading the disp Function on page 16-45
16
Custom Display Interface

In this section... Default Object Display on page 16-2 CustomDisplay Class on page 16-3 Methods for Customizing Object Display on page 16-3
Default Object Display

MATLAB adds default methods named disp and display to all MATLAB classes that do not implement their own methods with those names. These methods are not visible, but create the default simple display. The default simple display consists of the following parts: A header showing the class name, and the dimensions for nonscalar arrays. A list of all nonhidden public properties, shown in the order of definition in the class. The actual display depends on whether the object is scalar or nonscalar. Also, there are special displays for a scalar handle to a deleted object and empty object arrays. Object in all of these states are displayed differently if the objects have no properties. The details method creates the default detailed display. The detailed display adds these items to the simple display: Use of fully qualified class names Link to handle class, if the object is a handle Links to methods, events, and superclasses functions executed on the object. See Class with Default Object Display on page 16-15 for an example of how MATLAB displays objects.
16-2
Properties Displayed by Default

MATLAB displays object properties that have public get access and are not hidden (see Property Attributes on page 6-8). Inherited abstract properties are excluded from display. When the object being displayed is scalar, any dynamic properties attached to the object are also included.
CustomDisplay Class
The matlab.mixin.CustomDisplay class provides an interface that you can use to customize object display for your class. To use this interface, derive your class from CustomDisplay:
classdef MyClass < matlab.mixin.CustomDisplay
The CustomDisplay class is HandleCompatibile, so you can use it in combination with both value and handle superclasses.
disp, display, and details

The CustomDisplay class defines three sealed public methods. These methods overload three MATLAB functions disp, display, and details. The disp and display methods behave like the equivalent MATLAB functions, but use the customizations defined by classes derived from CustomDisplay. The details method always uses the default detailed object display and does not apply customizations defined for the class. The CustomDisplay interface does not allow you to override disp and display. Instead, you override any combination of the customization methods defined for this purpose.
Methods for Customizing Object Display

There are two groups of methods that you use to customize object display for your class: Part builder methods build the strings used for the standard display. Override any of these methods to change the respective parts of the display.
16-3
16
State handler methods are called for objects in specific states, like scalar, nonscalar, and so on. Override any of these methods to handle objects in a specific state. All of these methods have protected access and must be defined as protected in your subclass of CustomDisplay (that is, Access = protected).
Parts of an Object Display

There are three parts that makeup the standard object display header, property list, and footer For example, here is the standard object display for a containers.Map object:
The default object display does not include a footer. The detailed display provides more information:
16-4
Part Builder Methods

Each part of the object display has an associated method that assembles the respective part of the display. Method
getHeader
Purpose Create the string used for the header
Default Returns the string, [class(obj), ' with properties:'] linking the class name to a help popup Returns an array of PropertyGroup objects, which determines how to display the properties Simple display returns an empty string Detailed display returns linked calls to methods, events, and superclasses for this class
getPropertyGroups
Define how and what properties display, including order, values, and grouping Create the string used for the footer
getFooter
Object States That Affect Display

There are four object states that affect how MATLAB displays objects: Valid scalar object Nonscalar object array Empty object array Scalar handle to a deleted object
State Handler Methods

Each object state has an associated method that MATLAB calls whenever displaying objects that are in that particular state.
16-5
16
State Handler Method

displayScalarObject displayNonScalarObject displayEmptyObject displayScalarHandleToDeletedObject
Called for Object in This State

(isa(obj,'handle') && isvalid(obj)) && prod(size(obj)) == 1 prod(size(obj)) > 1 prod(size(obj)) == 0 isa(obj,'handle') && isscalar(obj) && ~isvalid(obj)
Utility Methods
The CustomDisplay class provides utility methods that return strings that are used in various parts of the different display options. These static methods return text that simplifies the creation of customized object displays. If the computer display does not support hypertext linking, the strings are returned without the links. Method
convertDimensionsToString
Inputs Valid object array
Outputs Object dimensions converted to a string; determined by calling size(obj) Displays the titles and property groups defined
displayPropertyGroups
PropertyGroup array
getClassNameForHeader getDeletedHandleText
Object None
Simple class name linked to the objects documentation String'handle to deleted' linked to the documentation on deleted handles String containing phrase
'Methods, Events, Superclasses', with each
getDetailedFooter
Object
link executing the respective command on the input object
16-6
Method
getDetailedHeader
Inputs Object
Outputs String containing linked class name, link to handle page (if handle class) and 'with
properties:'
getHandleText
None
String 'handle' linked to a section of the documentation that describes handle objects String containing linked class name and the phrase 'with
properties:'
getSimpleHeader
Object
16-7
16
How CustomDisplay Works

In this section... Steps to Display an Object on page 16-8 Methods Called for a Given Object State on page 16-9
Steps to Display an Object

When displaying an object, MATLAB determines the state of the object and calls the appropriate method for that state (seeObject States That Affect Display on page 16-5). For example, suppose obj is a valid scalar object of a class derived from CustomDisplay. If you type obj at the command line without terminating the statement with a semicolon:
>> obj
The following sequence results in the display of obj:

1 MATLAB determines the class of obj and calls the disp method to display
the object.
2 disp calls size to determine if obj is scalar or nonscalar 3 When obj is a scalar handle object, disp calls isvalid to determine if obj
is the handle of a deleted object. Deleted handles in nonscalar arrays do not affect the display.
4 disp calls the state handler method for an object of the state of obj. In this
case, obj is a valid scalar that results in a call to:

displayScalarObject(obj)
5 displayScalarObject calls the display part-builder methods to provide
the respective header, property list, and footer.

... header = getHeader(obj); disp(header)
16-8
How CustomDisplay Works
... groups = getPropertyGroups(obj) displayPropertyGroups(obj,groups) ... footer = getFooter disp(footer)
MATLAB follows a similar sequence for nonscalar object arrays, empty object arrays. In the case of scalar handles to deleted objects, disp calls the displayScalarHandleToDeletedObject method, which displays the default string for handles to deleted objects without calling any part-builder methods.
Methods Called for a Given Object State

The following diagram illustrates the methods called to display an object that derives from CustomDisplay. The disp method calls the state handler method that is appropriate for the state of the object or object array being displayed. Only an instance of a handle class can be in a state of scalar handle to a deleted object.
16-9
16
16-10
Role of the size Function In Custom Displays
Role of the size Function In Custom Displays

In this section... How size Is Used on page 16-11 Precautions When Overloading size on page 16-11
How size Is Used

In the process of building the custom display, CustomDisplay methods call the size function at several points: disp calls size to determine which state handler method to invoke The default getHeader method calls size to determine whether to display a scalar or nonscalar header The default displayPropertyGroups method calls size to determine if it should look up property values when the property group is a cell array of property names. By default, only scalar objects display the values of properties.
Precautions When Overloading size

If your class overloads the size function, then MATLAB calls the overloading version. You must ensure that the implementation of size is consistent with the way you want to display objects of the class. An unusual or improper implementation of size can result in undesirable display behavior. For example, suppose a class overloads size reports an object as scalar when it is not. In this class, a property list consisting of a cell array of strings results in the property values of the first object of the array being displayed. This behavior can give the impression that all objects in the array have the same property values. However, reporting an object as scalar when in fact the object is empty results in the object displaying as an empty object array. The default methods of the CustomDisplay interface always determine if the input is an empty array before attempting to access property values.
16-11
16
As you override CustomDisplay methods to implement your custom object display, consider how an overloading size method can affect the result.
16-12
Custom Display for Heterogeneous Arrays
Custom Display for Heterogeneous Arrays

You can call only sealed methods on nonscalar heterogeneous arrays. If you want to customize classes that are part of a heterogeneous hierarchy, you must override and declare as Sealed all the methods that are part of the CustomDisplay interface. The versions of disp and display that are inherited from matlab.mixin.CustomDisplay are sealed. However, these methods call all of the part builder (Part Builder Methods on page 16-5) and state handler methods (State Handler Methods on page 16-5). To use the CustomDisplay interface, the root class of the heterogeneous hierarchy can declare these methods as Sealed and Access = protected. If you do not need to override a particular method, then call the superclass method, as shown in the following code. For example, the following code shows modifications to the getPropertyGroups and displayScalarObject methods, while using the superclass implementation of all others.
classdef RootClass < matlab.mixin.CustomDisplay & matlab.mixin.Heterogeneous %... methods (Sealed, Access = protected) function header = getHeader(obj) header = [email protected](obj); end function groups = getPropertyGroups(obj) % Override of this method % ... end function footer = getFooter(obj) footer = [email protected](obj); end function displayNonScalarObject(obj) [email protected](obj);
16-13
16
end function displayScalarObject(obj) % Override of this method % ... end function displayEmptyObject(obj) [email protected](obj); end function displayScalarHandleToDeletedObject(obj) [email protected](obj); end end end
You do not need to declare the inherited static methods as Sealed.
16-14
Class with Default Object Display

In this section... The EmployeeInfo Class on page 16-15 Default Display Scalar on page 16-16 Default Display Nonscalar on page 16-17 Default Display Empty Object Array on page 16-17 Default Display Handle to Deleted Object on page 16-18 Default Display Detailed Display on page 16-18
The EmployeeInfo Class

The EmployeeInfo class defines a number of properties to store information about company employees. This simple class serves as the example class used in display customization sample classes.
EmployeeInfo derives from the matlab.mixin.CustomDisplay class to enable
customization of the object display.

EmployeeInfo is also a handle class. Therefore instances of this class can be
in the state referred to as a handle to a deleted object. This state does not occur with value classes (classes not derived for handle).
classdef EmployeeInfo < handle & matlab.mixin.CustomDisplay properties Name JobTitle Department Salary Password end methods function obj = EmployeeInfo obj.Name = input('Name: '); obj.JobTitle = input('Job Title: '); obj.Department = input('Department: '); obj.Salary = input('Salary: ');
16-15
16
obj.Password end end end
= input('Password: ');
The matlab.mixin.CustomDisplay is handle compatible. Therefore, superclasses can be either handle or value classes.
Default Display Scalar

Here is the creation and display of a scalar EmployeeInfo object. By default, MATLAB displays properties and their values for scalar objects. Provide inputs for the constructor:
>>Emp123 = EmployeeInfo; Name: 'Bill Tork' Job Title: 'Software Engineer' Department: 'Product Development' Salary: 1000 Password: 'bill123'
Display the object:

>>Emp123 Emp123 = EmployeeInfo with properties: Name: JobTitle: Department: Salary: Password: 'Bill Tork' 'Software Engineer' 'Product Development' 1000 'bill123'
Testing for Scalar Objects

To test for scalar objects, use isscalar.
16-16
Default Display Nonscalar

The default display for an array of objects does not show property values. For example, concatenating two EmployeeInfo objects generates this display:
>>[Emp123,Emp124] ans 1x2 EmployeeInfo array with properties: Name JobTitle Department Salary Password
Testing for Nonscalar Objects

To test for nonscalar objects, use a negated call to isscalar .
Default Display Empty Object Array

An empty object array has at least one dimension equal to zero.
>> Empt = EmployeeInfo.empty(0,5) Empt = 0x5 EmployeeInfo array with properties: Name JobTitle Department Salary Password
Testing for Empty Object Arrays

Use isempty to test for empty object arrays. An empty object array is not scalar because its dimensions can never be 1by-1.
>> emt = EmployeeInfo.empty
16-17
16
emt = 0x0 EmployeeInfo array with properties: Name JobTitle Department Salary Password >> isscalar(emt) ans = 0
Default Display Handle to Deleted Object

When a handle object is deleted, the handle variable can remain in the workspace.
>> delete(Emp123) >> Emp123 Emp123 = handle to deleted EmployeeInfo
Testing for Handles to Deleted Objects

To test for a handle to a deleted object, use isvalid. Note isvalid is a handle class method. Calling isvalid on a value class object causes an error.
Default Display Detailed Display

The details method does not support customization and always returns the standard detailed display:
details(Emp123)
16-18
EmployeeInfo handle with properties: Name: JobTitle: Department: Salary: Password: 'Bill Tork' 'Software Engineer' 'Product Development' 1000 'bill123'
Methods, Events, Superclasses
16-19
16
How to Approach Display Customization

In this section... Ways to Implement a Custom Display on page 16-20 Sample Approaches Using the Interface on page 16-21
Ways to Implement a Custom Display

How you customize object display for your class depends on what parts of the display you want to customize and what object states you want to use a custom display. In general, if you are making small changes to the default layout, then override the relevant part builder methods (Part Builder Methods on page 16-5). For example, suppose you want to: Change the order or value of properties, display a subset of properties, or create property groups Modify the header string Add a footer If you are defining a nonstandard display for a particular object state (scalar, for example), then the best approach is to override the appropriate state handler method (State Handler Methods on page 16-5). In some cases, a combination of method overrides might be the best approach. For example, your implementation of displayScalarObject might Use some of the utility methods (Utility Methods on page 16-6) to build your own display strings using parts from the default display Call a part builder method to get the default string for that particular part of the display Implement a completely different display for scalar objects. Once you override any CustomDisplay method, your override is called in all cases where the superclass method would have been called. For example, if
16-20
you override the getHeader method, your override must handle all cases where a state handler method calls getHeader. (See Methods Called for a Given Object State on page 16-9)
Sample Approaches Using the Interface

Here are some simple cases that show what methods to use for the particular customized display.
Change the Display of Scalar Objects

Use a nonstandard layout for scalar object display that is fully defined in the displayScalarObject method:
classdef MyClass < matlab.mixin.CustomDisplay ... methods (Access = protected) function displayScalarObject(obj) % Implement the custom display for scalar obj end end end
Custom Property List with Standard Layout

Use standard display layout, but create a custom property list for scalar and nonscalar display:
classdef MyClass < matlab.mixin.CustomDisplay ... methods(Access = protected) function groups = getPropertyGroups(obj) % Return PropertyGroup instances end end end
16-21
16
Custom Property List for Scalar Only

Use standard display layout, but create a custom property list for scalar only. Call the superclass getPropertyGroups for the nonscalar case.
classdef MyClass < matlab.mixin.CustomDisplay properties Prop1 Prop2 Prop3 end methods(Access = protected) function groups = getPropertyGroups(obj) if isscalar(obj) % Scalar case: change order propList = {'Prop2','Prop1,'Prop3'}; groups = matlab.mixin.util.PropertyGroup(propList) else % Nonscalar case: call superclass method groups = [email protected](obj); end end end end
Custom Property List with Modified Values

Change the values displayed for some properties in the scalar case by creating property/value pairs in a struct. This getPropertyGroups method displays only Prop1 and Prop2, and displays the value of Prop2 as Prop1 divided by Prop3.
classdef MyClass < matlab.mixin.CustomDisplay properties Prop1 Prop2 Prop3 end methods(Access = protected) function groups = getPropertyGroups(obj)
16-22
if isscalar(obj) % Specify the values to be displayed for properties propList = struct('Prop1',obj.Prop1,... 'Prop2',obj.Prop1/obj.Prop3); groups = matlab.mixin.util.PropertyGroup(propList) else % Nonscalar case: call superclass method groups = [email protected](obj); end end end end
Complete Class Definitions

For complete class implementations, see these sections: Customize Property Display on page 16-24 Customize Header, Property List, and Footer on page 16-28 Customize Display of Scalar Objects on page 16-35 Custom Display of Object Arrays on page 16-40
16-23
16
Customize Property Display

In this section... Change the Property Order on page 16-24 Change the Values Displayed for Properties on page 16-25
Change the Property Order

This example shows how to change the order and number of properties displayed for an object of your class. Suppose your class definition contains the following property definition:
properties Name JobTitle Department Salary Password end
In the default scalar object display, MATLAB displays all the public properties along with their values. However, you want to display only Department, JobTitle, and Name, in that order. You can do this by deriving from CustomDisplay and overriding the getPropertyGroups method. Your override Defines method Access as protected to match the definition in the CustomDisplay superclass Creates a cell array of property names in the desired order Returns a PropertyGroup object constructed from the property list cell array
methods (Access = protected) function propgrp = getPropertyGroups(~) proplist = {'Department','JobTitle','Name'}; propgrp = matlab.mixin.util.PropertyGroup(proplist); end
16-24
end
When you create a PropertyGroup object using a cell array of property names, MATLAB automatically Adds the property values for a scalar object display Uses the property names without values for a nonscalar object display (including empty object arrays) The getPropertyGroups method is not called to create the display for a scalar handle to a deleted object.
Change the Values Displayed for Properties

Given the same class properties used in the previous section, you can change the value displayed for properties by building the property list as a struct and specifying values for property names. This override of the getPropertyGroups method uses the default property display for nonscalar objects by calling the superclass getPropertyGroups method. For scalar objects, the override: Changes the value displayed for the Password property to a '*' character for each character in the password string. Displays the string 'Not Available' for the Salary property.
methods (Access = protected) function propgrp = getPropertyGroups(obj) if ~isscalar(obj) propgrp = [email protected](obj); else pd(1:length(obj.Password)) = '*'; propList = struct('Department',obj.Department,... 'JobTitle',obj.JobTitle,... 'Name',obj.Name,... 'Salary','Not avalable',... 'Password',pd); propgrp = matlab.mixin.util.PropertyGroup(propList); end end end
16-25
16
The object display looks like this:

EmployeeInfo with properties: Department: JobTitle: Name: Salary: Password: 'Product Development' 'Software Engineer' 'Bill Tork' 'Not avalable' '*******'
Full Class Listing

classdef EmployeeInfo < handle & matlab.mixin.CustomDisplay properties Name JobTitle Department Salary Password end methods function obj = EmployeeInfo obj.Name obj.JobTitle obj.Salary obj.Password end end methods (Access = protected) function propgrp = getPropertyGroups(obj) if ~isscalar(obj) propgrp = [email protected](obj); else pd(1:length(obj.Password)) = '*'; propList = struct('Department',obj.Department,... 'JobTitle',obj.JobTitle,... 'Name',obj.Name,... 'Salary','Not avalable',... 'Password',pd); propgrp = matlab.mixin.util.PropertyGroup(propList); = input('Name: '); = input('Job Title: '); = input('Salary: '); = input('Password: ');
obj.Department = input('Department: ');
16-26
end end end
16-27
16
Customize Header, Property List, and Footer

In this section... Design of Custom Display on page 16-28 getHeader Method Override on page 16-30 getPropertyGroups Override on page 16-31 getFooter Override on page 16-32
Design of Custom Display

Note This example uses the EmployeeInfo class described in the Class with Default Object Display on page 16-15 section. This example shows how to customize each of the three parts of the display. For the header: Use default header for nonscalar object arrays. Build header string with linked class name and department name (from Department property) For properties: Nonscalar object arrays display a subset of property names in a different order than the default. Scalar objects create two property groups that have titles (Public Info and Personal Info). For the footer: Add a footer to the display, only when the object is a valid scalar that displays property values. Here is the customized display of an object of the EmployeeInfo class.
16-28
Emp123 = EmployeeInfo Dept: Product Development Public Info Name: 'Bill Tork' JobTitle: 'Software Engineer' Personal Info Salary: 1000 Password: 'bill123' Company Private
Here is the custom display of an array of EmployeeInfo objects:

[Emp123,Emp124] ans = 1x2 EmployeeInfo array with properties: Department Name JobTitle
Here is the display of an empty object array:

>> EmployeeInfo.empty(0,5) ans = 0x5 EmployeeInfo array with properties: Department Name JobTitle
Here is the display of a handle to a delete object (EmployeeInfo is a handle class):
16-29
16
>> delete(Emp123) >> Emp123 Emp123 = handle to deleted EmployeeInfo
Implementation
The EmployeeInfo class overrides three matlab.mixin.CustomDisplay methods to implement the display shown: getHeader getPropertyGroups getFooter Each method must produce the desired results with each of the following inputs: Scalar object Nonscalar object array Empty object array
getHeader Method Override

MATLAB calls getHeader to get the header string. The EmployeeInfo class overrides this method to implement the custom header for scalar display. Here is how it works: Nonscalar (including empty object) arrays call the superclass getHeader, which returns the default header. Scalar handles to deleted objects do not result in a call to getHeader. Scalar inputs build a custom header using the getClassNameForHeader static method to return a linked class name string, and the value of the Department property.
16-30
Here is the EmployeeInfo override of the getHeader method. The required protected access is inherited from the superclass.
methods (Access = protected) function header = getHeader(obj) if ~isscalar(obj) header = [email protected](obj); else className = matlab.mixin.CustomDisplay.getClassNameForHeader(obj); newHeader = [className,' Dept: ',obj.Department]; header = sprintf('%s\n',newHeader); end end end
getPropertyGroups Override
MATLAB calls getPropertyGroups to get the PropertyGroup objects, which control how properties are displayed. This method override defines two different property lists depending on the objects state: For nonscalar inputs, including empty arrays and arrays containing handles to deleted objects, create a property list as a cell array to reorder properties. By default, MATLAB does not display property values for nonscalar inputs. For scalar inputs, create two property groups with titles. The scalar code branch lists properties in a different order than the nonscalar case and includes Salary and Password properties. MATLAB automatically assigns property values. Scalar handles to deleted objects do not result in a call to getPropertyGroups. Both branches return a matlab.mixin.util.PropertyGroup object, which determines how to displays the object properties. Here is the EmployeeInfo override of the getPropertyGroups method. The protected access is inherited from the superclass.
methods (Access = protected) function propgrp = getPropertyGroups(obj)
16-31
16
if ~isscalar(obj) propList = {'Department','Name','JobTitle'}; propgrp = matlab.mixin.util.PropertyGroup(propList); else gTitle1 = 'Public Info'; gTitle2 = 'Personal Info'; propList1 = {'Name','JobTitle'}; propList2 = {'Salary','Password'}; propgrp(1) = matlab.mixin.util.PropertyGroup(propList1,gTitle1); propgrp(2) = matlab.mixin.util.PropertyGroup(propList2,gTitle2); end end end
getFooter Override
MATLAB calls getFooter to get the footer string. The EmployeeInfo getFooter method defines a footer for the display, which is included only when the input is a valid scalar object. In all other cases, getFooter returns an empty string. Scalar handles to deleted objects do not result in a call to getFooter.
methods (Access = protected) function footer = getFooter(obj) if isscalar(obj) footer = sprintf('%s\n','Company Private'); else footer = ''; end end end
Complete Class Listing

classdef EmployeeInfo < handle & matlab.mixin.CustomDisplay properties Name JobTitle Department Salary
16-32
Password end methods function obj = EmployeeInfo obj.Name obj.JobTitle obj.Salary obj.Password end end methods (Access = protected) function header = getHeader(obj) if ~isscalar(obj) header = [email protected](obj); else className = matlab.mixin.CustomDisplay.getClassNameForHeader(obj); newHeader = [className,' Dept: ',obj.Department]; header = sprintf('%s\n',newHeader); end end function propgrp = getPropertyGroups(obj) if ~isscalar(obj) propList = {'Department','Name','JobTitle'}; propgrp = matlab.mixin.util.PropertyGroup(propList); else gTitle1 = 'Public Info'; gTitle2 = 'Personal Info'; propList1 = {'Name','JobTitle'}; propList2 = {'Salary','Password'}; propgrp(1) = matlab.mixin.util.PropertyGroup(propList1,gTitle1); propgrp(2) = matlab.mixin.util.PropertyGroup(propList2,gTitle2); end end function footer = getFooter(obj) if isscalar(obj) footer = sprintf('%s\n','Company Private'); = input('Name: '); = input('Job Title: '); = input('Salary: '); = input('Password: ');
16-33
16
else footer = ''; end end end end
16-34
Customize Display of Scalar Objects

In this section... Design Of Custom Display on page 16-35 displayScalarObject Method Override on page 16-36 getPropertyGroups Override on page 16-37
Design Of Custom Display

Note This example uses the EmployeeInfo class described in the Class with Default Object Display on page 16-15 section. This example shows how to customize the display of scalar objects. The objective of this customized display is to: Modify the header to include the department name obtained from the Department property Group properties into two categories titled Public Info and Personal Info. Modify which properties are displayed Modify the values displayed for Personal Info category Use the default displayed for nonscalar objects, including empty arrays, and scalar deleted handles For example, here is the customized display of an object of the EmployeeInfo class.
Emp123 = EmployeeInfo Dept: Product Development Public Info Name: 'Bill Tork' JobTitle: 'Software Engineer'
16-35
16
Personal Info Salary: 'Level: 10' Password: '*******'
Implementation
The EmployeeInfo class overrides two matlab.mixin.CustomDisplay methods to implement the display shown: displayScalarObject Called to display valid scalar objects getPropertyGroups Builds the property groups for display
displayScalarObject Method Override

MATLAB calls displayScalarObject to display scalar objects. The EmployeeInfo class overrides this method to implement the scalar display. Once overridden, this method must control all aspects of scalar object display, including creating the header, property groups, and footer, if used. This implementation: Builds a custom header using the getClassNameForHeader static method to return a linked class name string and the value of the Department property to get the department name. Uses
sprintf to add a new line to the header string
Displays the header with the built-in disp function. Calls the getPropertyGroups override to define the property groups (see following section). Displays the property groups using the displayPropertyGroups static method. Here is the EmployeeInfo override of the displayScalarObject method. The required protected access is inherited from the superclass.
methods (Access = protected) function displayScalarObject(obj)
16-36
className = matlab.mixin.CustomDisplay.getClassNameForHeader(obj); scalarHeader = [className,' Dept: ',obj.Department]; header = sprintf('%s\n',scalarHeader); disp(header) propgroup = getPropertyGroups(obj); matlab.mixin.CustomDisplay.displayPropertyGroups(obj,propgroup) end end
getPropertyGroups Override
MATLAB calls getPropertyGroups when displaying scalar or nonscalar objects. However, MATLAB does not call this method when displaying a scalar handle to a deleted object. The EmployeeInfo class overrides this method to implement the property groups for scalar object display. This implementation calls the superclass getPropertyGroups method if the input is not scalar. If the input is scalar, this method: Defines two titles for the two groups Creates a cell array of property names that are included in the first group. MATLAB adds the property values for the display Creates a struct array of property names with associated property values for the second group. Using a struct instead of a cell array enables you to replace the values that are displayed for the Salary and Password properties without changing the personal information stored in the object properties. Constructs two matlab.mixin.util.PropertyGroup objects, which are used by the displayScalarObject method. Here is the EmployeeInfo override of the getPropertyGroups method. The required protected access is inherited from the superclass.
methods (Access = protected) function propgrp = getPropertyGroups(obj) if ~isscalar(obj) propgrp = [email protected](obj);
16-37
16
else gTitle1 = 'Public Info'; gTitle2 = 'Personal Info'; propList1 = {'Name','JobTitle'}; pd(1:length(obj.Password)) = '*'; level = round(obj.Salary/100); propList2 = struct('Salary',... ['Level: ',num2str(level)],... 'Password',pd); propgrp(1) = matlab.mixin.util.PropertyGroup(propList1,gTitle1); propgrp(2) = matlab.mixin.util.PropertyGroup(propList2,gTitle2); end end end

classdef EmployeeInfo4 < handle & matlab.mixin.CustomDisplay properties Name JobTitle Department Salary Password end methods function obj = EmployeeInfo4 obj.Name obj.JobTitle obj.Salary obj.Password end end methods (Access = protected) function displayScalarObject(obj) className = matlab.mixin.CustomDisplay.getClassNameForHeader(obj); scalarHeader = [className,' Dept: ',obj.Department]; header = sprintf('%s\n',scalarHeader); = input('Name: '); = input('Job Title: '); = input('Salary: '); = input('Password: ');
16-38
disp(header) propgroup = getPropertyGroups(obj); matlab.mixin.CustomDisplay.displayPropertyGroups(obj,propgroup) end function propgrp = getPropertyGroups(obj) if ~isscalar(obj) propgrp = [email protected](obj); else % property groups for scalars gTitle1 = 'Public Info'; gTitle2 = 'Personal Info'; propList1 = {'Name','JobTitle'}; pd(1:length(obj.Password)) = '*'; level = round(obj.Salary/100); propList2 = struct('Salary',... ['Level: ',num2str(level)],... 'Password',pd); propgrp(1) = matlab.mixin.util.PropertyGroup(propList1,gTitle1); propgrp(2) = matlab.mixin.util.PropertyGroup(propList2,gTitle2); end end end end
16-39
16
Custom Display of Object Arrays

In this section... Design of Custom Display on page 16-40 The displayNonScalarObject Override on page 16-41 The displayEmptyObject Override on page 16-42
Design of Custom Display

Note This example uses the EmployeeInfo class described in the Class with Default Object Display on page 16-15 section. This example shows how to customize the display of nonscalar objects, including empty object arrays. The objective of this customized display is to: Construct a custom header using some elements of the default header Display a subset of property-specific information for each object in the array. List handles to deleted objects in the array using a string with links to documentation for handle objects and the class. Display empty objects with a slight modification to the default header Here is the customized display of an array of three EmployeeInfo objects
1x3 EmployeeInfo array with members: 1. Employee: Name: 'Bill Tork' Department: 'Product Development' 2. Employee: Name: 'Alice Blackwell' Department: 'QE' 3. Employee: Name: 'Nancy Green'
16-40
Department: 'Documentation'
Deleted object handles in the array indicate their state:

1x3 EmployeeInfo members: 1. Employee: Name: 'Bill Tork' Department: 'Product Development' 2. handle to deleted EmployeeInfo 3. Employee: Name: 'Nancy Green' Department: 'Documentation'
To achieve the desired result, the EmployeeInfo class overrides the following methods of the matlab.mixin.CustomDisplay class: displayNonScalarObjects Called to display nonempty object arrays displayEmptyObjects Called to display empty object arrays
The displayNonScalarObject Override

MATLAB calls the displayNonScalarObject method to display object arrays. The override of this method in the EmployeeInfo class: Builds a header string using convertDimensionsToString to obtain the array size and getClassNameForHeader to get the class name with a link to the help for that class. Displays the modified header string. Loops through the elements in the array, building two different subheaders depending on the individual object state. In the loop, this method:
Detects handles to deleted objects (using the isvalid handle class method). Uses getDeletedHandleText and getClassNameForHeader to build a string for array elements that are handles to deleted objects. Builds a custom subheader for valid object elements in the array
16-41
16
Creates a PropertyGroup object containing the Name and Department properties for valid objects Uses the displayPropertyGroups static method to generate the property display for valid objects. Here is the implementation of displayNonScalarObjects:
methods (Access = protected) function displayNonScalarObject(objAry) dimStr = matlab.mixin.CustomDisplay.convertDimensionsToString(objAry); cName = matlab.mixin.CustomDisplay.getClassNameForHeader(objAry); headerStr = [dimStr,' ',cName,' members:']; header = sprintf('%s\n',headerStr); disp(header) for ix = 1:length(objAry) o = objAry(ix); if ~isvalid(o) str1 = matlab.mixin.CustomDisplay.getDeletedHandleText; str2 = matlab.mixin.CustomDisplay.getClassNameForHeader(o); headerInv = [str1,' ',str2]; tmpStr = [num2str(ix),'. ',headerInv]; numStr = sprintf('%s\n',tmpStr); disp(numStr) else numStr = [num2str(ix),'. Employee:']; disp(numStr) propList = struct('Name',o.Name,... 'Department',o.Department); propgrp = matlab.mixin.util.PropertyGroup(propList); matlab.mixin.CustomDisplay.displayPropertyGroups(o,propgrp); end end end end
The displayEmptyObject Override

MATLAB calls the displayEmptyObject method to display empty object arrays. The implementation of this method in the EmployeeInfo class builds a custom header for empty objects following these steps:
16-42
Gets the array dimensions in string format using the convertDimensionsToString static method. Gets a string with the class name linked to the helpPopup function using the getClassNameForHeader static method. Builds and displays the custom string for empty arrays.
methods (Access = protected) function displayEmptyObject(obj) dimstr = matlab.mixin.CustomDisplay.convertDimensionsToString(obj); className = matlab.mixin.CustomDisplay.getClassNameForHeader(obj); emptyHeader = [dimstr,' ',className,' with no employee information']; header = sprintf('%s\n',emptyHeader); disp(header) end end
For example, an empty EmployeeInfo object displays like this:

Empt = EmployeeInfo.empty(0,5) Empt = 0x5 EmployeeInfo with no employee information

classdef EmployeeInfo < handle & matlab.mixin.CustomDisplay properties Name JobTitle Department Salary Password end methods function obj = EmployeeInfo obj.Name obj.JobTitle = input('Name: '); = input('Job Title: ');
16-43
16
obj.Salary obj.Password end end
= input('Salary: '); = input('Password: ');
methods (Access = protected) function displayNonScalarObject(objAry) dimStr = matlab.mixin.CustomDisplay.convertDimensionsToString(objAry); cName = matlab.mixin.CustomDisplay.getClassNameForHeader(objAry); headerStr = [dimStr,' ',cName,' members:']; header = sprintf('%s\n',headerStr); disp(header) for ix = 1:length(objAry) o = objAry(ix); if ~isvalid(o) str1 = matlab.mixin.CustomDisplay.getDeletedHandleText; str2 = matlab.mixin.CustomDisplay.getClassNameForHeader(o); headerInv = [str1,' ',str2]; tmpStr = [num2str(ix),'. ',headerInv]; numStr = sprintf('%s\n',tmpStr); disp(numStr) else numStr = [num2str(ix),'. Employee']; disp(numStr) propList = struct('Name',o.Name,... 'Department',o.Department); propgrp = matlab.mixin.util.PropertyGroup(propList); matlab.mixin.CustomDisplay.displayPropertyGroups(o,propgrp); end end end function displayEmptyObject(obj) dimstr = matlab.mixin.CustomDisplay.convertDimensionsToString(obj); className = matlab.mixin.CustomDisplay.getClassNameForHeader(obj); emptyHeader = [dimstr,' ',className,' with no employee information']; header = sprintf('%s\n',emptyHeader); disp(header) end end end
16-44
Overloading the disp Function
Overloading the disp Function

In this section... Display Methods on page 16-45 Implement disp Or disp and display on page 16-45 Relationship Between disp and display on page 16-46
Display Methods
Subclassing matlab.mixin.CustomDisplay is the best approach to customizing object display. However, if you do not derive your class from CustomDisplay, you can overload the disp function to change how MATLAB displays objects of your class. MATLAB calls an objects display method whenever an object is referred to in a statement that is not terminated by a semicolon. For example, the following statement creates the variable a. MATLAB calls display, which displays the value of a in the command line.
a = 5 a = 5 display then calls disp.
Implement disp Or disp and display

The built-in display function prints the name of the variable that is being displayed, if an assignment is made, or otherwise uses ans as the variable name. display then calls disp to handle the actual display of the values. If the variable that is being displayed is an object of a class that overloads disp, then MATLAB always calls the overloaded method. Overload disp or disp and display to customize the display of objects. Overloading only display is not sufficient to properly implement a custom display for your class.
16-45
16
Relationship Between disp and display

MATLAB invokes the built-in display function when: MATLAB executes a statement that returns a value and is not terminated with a semicolon. Code explicitly invokes the display function. MATLAB invokes the built-in disp function when: The built-in display function calls disp. Code explicitly invokes disp.
16-46
17
Implementing a Class for Polynomials
17
A Polynomial Class
In this section... Designing a Polynomial Object on page 17-2 Displaying the Class Files on page 17-2 Summary of the DocPolynom Class on page 17-3 The DocPolynom Constructor Method on page 17-4 Removing Irrelevant Coefficients on page 17-5 Converting DocPolynom Objects to Other Types on page 17-6 The DocPolynom disp Method on page 17-9 The DocPolynom subsref Method on page 17-10 Defining Arithmetic Operators for DocPolynom on page 17-13
Designing a Polynomial Object

This example implements a class to represent polynomials in the MATLAB language. The design requirements are: Value class behaviora polynomial object should behave like MATLAB numeric variables when copied and passed to functions. Specialized display and indexing Arithmetic operations Double converter simplifying the use of polynomial object with existing MATLAB functions that accept numeric inputs.

Open the DocPolynom class definition file in the MATLAB editor. To use the class, create a folder named @DocPolynom and save DocPolynom.m to this folder. The parent folder of @DocPolynom must be on the MATLAB path.
17-2
A Polynomial Class
Summary of the DocPolynom Class

The class definition specifies a property for data storage and defines a folder (@DocPolynom) that contains the class definition. The following table summarizes the properties defined for the DocPolynom class. DocPolynom Class Properties Name
coef
Class
double
Default
[]
Description Vector of polynomial coefficients [highest order ... lowest order]
The following table summarizes the methods for the DocPolynom class. DocPolynom Class Methods Name
DocPolynom double char disp subsref
Description Class constructor Converts a DocPolynom object to a double (i.e., returns its coefficients in a vector) Creates a formatted display of the DocPolynom object as powers of x and is used by the disp method Determines how MATLAB displays a DocPolynom objects on the command line Enables you to specify a value for the independent variable as a subscript, access the coef property with dot notation, and call methods with dot notation. Implements addition of DocPolynom objects Implements subtraction of DocPolynom objects Implements multiplication of DocPolynom objects
plus minus mtimes
17-3
17
Using the DocPolynom Class

The following examples illustrate basic use of the DocPolynom class. Create DocPolynom objects to represent the following polynomials. Note that the argument to the constructor function contains the polynomial coefficients and
p1 = DocPolynom([1 0 -2 -5]) p1 = x^3 - 2*x - 5 p2 = DocPolynom([2 0 3 2 -7]) p2 = 2*x^4 + 3*x^2 + 2*x - 7
Find the roots of the polynomial by passing the coefficients to the roots function.
roots(p1.coef) ans = 2.0946 -1.0473 + 1.1359i -1.0473 - 1.1359i
Add the two polynomials p1 and p2. MATLAB calls the plus method defined for the DocPolynom class when you add two DocPolynom objects.
p1 + p2 ans = 2*x^4 + x^3 + 3*x^2 - 12
The DocPolynom Constructor Method

The following function is the DocPolynom class constructor, which is in the file @DocPolynom/DocPolynom.m:
function obj = DocPolynom(c) % Construct a DocPolynom object using the coefficients supplied
17-4
A Polynomial Class
if isa(c,'DocPolynom') obj.coef = c.coef; else obj.coef = c(:).'; end end
Constructor Calling Syntax

It is possible to all the DocPolynom constructor with two different arguments: Input argument is a DocPolynom object If you call the constructor function with an input argument that is already a DocPolynom object, the constructor returns a new DocPolynom object with the same coefficients as the input argument. The isa function checks for this input. Input argument is a coefficient vector If the input argument is not a DocPolynom object, the constructor attempts to reshape the values into a vector and assign them to the coef property. The coef property set method restricts property values to doubles. See Removing Irrelevant Coefficients on page 17-5 for a description of the property set method. An example use of the DocPolynom constructor is the statement:
p = DocPolynom([1 0 -2 -5]) p = x^3 - 2*x -5
This statement creates an instance of the DocPolynom class with the specified coefficients. Note the display of the object shows the equivalent polynomial using MATLAB language syntax. The DocPolynom class implements this display using the disp and char class methods.
Removing Irrelevant Coefficients

MATLAB software represents polynomials as row vectors containing coefficients ordered by descending powers. Zeros in the coefficient vector
17-5
17
represent terms that drop out of the polynomial. Leading zeros, therefore, can be ignored when forming the polynomial. Some DocPolynom class methods use the length of the coefficient vector to determine the degree of the polynomial. It is useful, therefore, to remove leading zeros from the coefficient vector so that its length represents the true value. The DocPolynom class stores the coefficient vector in a property that uses a set method to remove leading zeros from the specified coefficients before setting the property value.
function obj = set.coef(obj,val) % coef set method if ~isa(val,'double') error('Coefficients must be of class double') end ind = find(val(:).'~=0); if ~isempty(ind); obj.coef = val(ind(1):end); else obj.coef = val; end end
Converting DocPolynom Objects to Other Types

The DocPolynom class defines two methods to convert DocPolynom objects to other classes: double Converts to the double numeric type so functions can perform mathematical operations on the coefficients. char Converts to string used to format output for display in the command window
The DocPolynom to Double Converter

The double converter method for the DocPolynom class simply returns the coefficient vector:
function c = double(obj)
17-6
A Polynomial Class
% DocPolynom/Double Converter c = obj.coef; end
For the DocPolynom object p:

p = DocPolynom([1 0 -2 -5]);
the statement:
c = double(p)
returns:
c= 1 0 -2 -5
which is of class double:

class(c) ans = double
The DocPolynom to Character Converter

The char method produces a character string that represents the polynomial displayed as powers of an independent variable, x. Therefore, after you have specified a value for x, the string returned is a syntactically correct MATLAB expression, which you can evaluate. The char method uses a cell array to collect the string components that make up the displayed polynomial. The disp method uses char to format the DocPolynom object for display. Class users are not likely to call the char or disp methods directly, but these methods enable the DocPolynom class to behave like other data classes in MATLAB. Here is the char method.
17-7
17
function str = char(obj) % Created a formated display of the polynom % as powers of x if all(obj.coef == 0) s = '0'; str = s; return else d = length(obj.coef)-1; s = cell(1,d); ind = 1; for a = obj.coef; if a ~= 0; if ind ~= 1 if a > 0 s(ind) = {' + '}; ind = ind + 1; else s(ind) = {' - '}; a = -a; ind = ind + 1; end end if a ~= 1 || d == 0 if a == -1 s(ind) = {'-'}; ind = ind + 1; else s(ind) = {num2str(a)}; ind = ind + 1; if d > 0 s(ind) = {'*'}; ind = ind + 1; end end end if d >= 2 s(ind) = {['x^' int2str(d)]}; ind = ind + 1; elseif d == 1
17-8
A Polynomial Class
s(ind) = {'x'}; ind = ind + 1; end end d = d - 1; end end str = [s{:}]; end
Evaluating the Output

If you create the DocPolynom object p:
p = DocPolynom([1 0 -2 -5]);
and then call the char method on p:

char(p)
the result is:

ans = x^3 - 2*x - 5
The value returned by char is a string that you can pass to eval after you have defined a scalar value for x. For example:
x = 3; eval(char(p)) ans = 16
The DocPolynom disp Method

To provide a more useful display of DocPolynom objects, this class overloads disp in the class definition. This disp method relies on the char method to produce a string representation of the polynomial, which it then displays on the screen.
17-9
17
function disp(obj) % DISP Display object in MATLAB syntax c = char(obj); % char returns a cell array if iscell(c) disp([' ' c{:}]) else disp(c) % all coefficients are zero end end
When MATLAB Calls the disp Method

The statement:
p = DocPolynom([1 0 -2 -5])
creates a DocPolynom object. Because the statement is not terminated with a semicolon, the resulting output is displayed on the command line:
p = x^3 - 2*x - 5
The DocPolynom subsref Method

Normally, subscripted assignment is defined automatically by MATLAB. However, in this particular case, the design of the DocPolynom class specifies that a subscripted reference to a DocPolynom object causes an evaluation of the polynomial with the value of the independent variable equal to the subscript. For example, given the following polynomial:
Create a DocPolynom object p:

p = DocPolynom([1 0 -2 -5]) p = x^3 - 2*x - 5
17-10
A Polynomial Class
The following subscripted expression evaluates the value of the polynomial at x = 3 and x = 4 and returns the resulting values:
p([3 4]) ans = 16 51
Defining the subsref Method

The DocPolynom class redefines the default subscripted reference behavior by implementing a subsref method. If a class defines a subsref method, MATLAB software calls this method for objects of this class whenever a subscripted reference occurs. You must, therefore, define all the behaviors you want your class to exhibit in the local method. The DocPolynom subsref method implements the following behaviors: The ability to pass a value for the independent variable as a subscripted reference (i.e., p(3) evaluates the polynomial at x = 3) Dot notation for accessing the coef property Dot notation for access to class methods, which accept and return differing numbers of input and output arguments
subsref Implementation Details

See the subsref function for general information on implementing this method. When you implement a subsref method to support calling methods with arguments using dot notation, both the type and subs structure fields contain multiple elements. For example, consider a call to the polyval function:
p = DocPolynom([1 0 -2 -5]) p = x^3 - 2*x - 5
17-11
17
polyval(p.coef,[3 5 7]) ans = 16 110 324
The polyval function requires an input argument of values at which to evaluate the polynomial and returns the value of f(x) at these values. subsref performs the method call through the statements:
case '()' ind = s.subs{:}; b = polyval(a.coef,ind);
The subsref method implements all subscripted reference explicitly, as show in the following code listing.
function b = subsref(a,s) switch s(1).type case '()' ind = s.subs{:}; b = polyval(a.coef,ind); case '.' switch s(1).subs case 'coef' b = a.coef; case 'disp' disp(a) otherwise if length(s)>1 b = a.(s(1).subs)(s(2).subs{:}); else b = a.(s.subs); end end otherwise error('Specify value for x as obj(x)') end end % subsref
17-12
A Polynomial Class
Defining Arithmetic Operators for DocPolynom

Several arithmetic operations are meaningful on polynomials and are implemented for the DocPolynom class. This section shows how to implement the following methods: Method and Syntax
plus(a,b) minus(a,b) mtimes(a,b)
Operator Implemented Addition Subtraction Matrix multiplication
When overloading arithmetic operators, keep in mind what data types you want to operate on. The plus, minus, and mtimes methods are defined for the DocPolynom class to handle addition, subtraction, and multiplication on DocPolynomDocPolynom and DocPolynomdouble combinations of operands.
Defining the + Operator

If either p or q is a DocPolynom object, the expression
p + q
generates a call to a function @DocPolynom/plus, unless the other object is of a class of higher precedence. The following function redefines the plus (+) operator for the DocPolynom class:
function r = plus(obj1,obj2) % Plus Implement obj1 + obj2 for DocPolynom obj1 = DocPolynom(obj1); obj2 = DocPolynom(obj2); k = length(obj2.coef) - length(obj1.coef); r = DocPolynom([zeros(1,k) obj1.coef]+[zeros(1,-k) obj2.coef]); end
Here is how the function works:
17-13
17
Ensure that both input arguments are DocPolynom objects so that expressions such as
p + 1
that involve both a DocPolynom and a double, work correctly. Access the two coefficient vectors and, if necessary, pad one of them with zeros to make both the same length. The actual addition is simply the vector sum of the two coefficient vectors. Call the DocPolynom constructor to create a properly typed result.
Defining the - Operator

You can implement the minus operator (-) using the same approach as the plus (+) operator. The MATLAB runtime calls the DocPolynom minus method to compute p - q, where p, q, or both are DocPolynom objects:
function r = minus(obj1,obj2) % MINUS Implement obj1 - obj2 for DocPolynom obj1 = DocPolynom(obj1); obj2 = DocPolynom(obj2); k = length(obj2.coef) - length(obj1.coef); r = DocPolynom([zeros(1,k) obj1.coef]-[zeros(1,-k) obj2.coef]); end
Defining the * Operator

MATLAB calls the DocPolynom mtimes method to compute the product p*q. The mtimes method implements matrix multiplication since the multiplication of two polynomials is simply the convolution (conv) of their coefficient vectors:
function r = mtimes(obj1,obj2) % MTIMES Implement obj1 * obj2 for DocPolynoms obj1 = DocPolynom(obj1); obj2 = DocPolynom(obj2); r = DocPolynom(conv(obj1.coef,obj2.coef)); end
17-14
A Polynomial Class
Using the Arithmetic Operators

Given the DocPolynom object:
p = DocPolynom([1 0 -2 -5]);
The following two arithmetic operations call the DocPolynom plus and mtimes methods:
q = p+1 r = p*q
to produce
q = x^3 - 2*x - 4 r = x^6 - 4*x^4 - 9*x^3 + 4*x^2 + 18*x + 20
17-15
17
17-16
18
Designing Related Classes
A Simple Class Hierarchy on page 18-2 Containing Assets in a Portfolio on page 18-19
18
A Simple Class Hierarchy

In this section... Shared and Specialized Properties on page 18-2 Designing a Class for Financial Assets on page 18-3 Displaying the Class Files on page 18-4 Summary of the DocAsset Class on page 18-4 The DocAsset Constructor Method on page 18-5 The DocAsset Display Method on page 18-6 Designing a Class for Stock Assets on page 18-7 Displaying the Class Files on page 18-7 Summary of the DocStock Class on page 18-7 Designing a Class for Bond Assets on page 18-10 Displaying the Class Files on page 18-10 Summary of the DocBond Class on page 18-11 Designing a Class for Savings Assets on page 18-15 Displaying the Class Files on page 18-15 Summary of the DocSavings Class on page 18-15
Shared and Specialized Properties

As an example of how subclasses are specializations of more general classes, consider an asset class that can be used to represent any item that has monetary value. Some examples of assets are stocks, bonds, and savings accounts. This example implements four classes DocAsset, and the subclasses DocStock, DocBond, DocSavings. The DocAsset class holds the data that is common to all of the specialized asset subclasses in class properties. The subclasses inherit the super class properties in addition to defining their own properties. The subclasses are all kinds of assets.
18-2
The following diagram shows the properties defined for the classes of assets.
DocAsset Properties Description Date Type CurrentValue
DocStocks (is a DocAsset) Inherited Props Description Date Type CurrentValue DocStocks Props NumShares SharePrice
DocBonds (is a DocAsset) Inherited Props Description Date Type CurrentValue DocBonds Props FaceValue Yield CurrentBondYield
DocSavings (is a DocAsset) Inherited Props Description Date Type CurrentValue DocSavings Props InterestRate
The DocStock, DocBond, and DocSavings classes inherit properties from the DocAsset class. In this example, the DocAsset class provides storage for data common to all subclasses and shares methods with these subclasses.
Designing a Class for Financial Assets

This class provides storage and access for information common to all asset children. It is not intended to be instantiated directly, so it does not require an extensive set of methods. The class contains the following methods: Constructor
18-3
18
A local setter function for one property

Open the DocAsset class definition file in the MATLAB Editor. To use the class, create a folder named @DocAsset and save DocAsset.m to this folder. The parent folder of @DocAsset must be on the MATLAB path.
Summary of the DocAsset Class

The class is defined in one file, DocAsset.m, which you must place in an @ folder of the same name. The parent folder of the @DocAsset folder must be on the MATLAB path. See the addpath function for more information. The following table summarizes the properties defined for the DocAsset class. DocAsset Class Properties Name
Description CurrentValue Date Type
Class
char double char char
Default
'' 0 date 'savings'
Description Description of asset Current value of asset Date when record is created (set by date function) Type of asset (stock, bond, savings)
The following table summarizes the methods for the DocAsset class. DocAsset Class Methods Name
DocAsset disp set.Type
Description Class constructor Displays information about this object Set function for Type. Property tests for correct value when property is set.
18-4
The DocAsset Constructor Method

This class has four properties that store data common to all of the asset subclasses. All except Date are passed to the constructor by a subclass constructor. Date is a private property and is set by a call to the date function. Description A character string that describes the particular asset (e.g., stock name, savings bank name, bond issuer, and so on). Date The date the object was created. This propertys set access is private so that only the constructor assigns the value using the date command when creating the object. Type The type of asset (e.g., savings, bond, stock). A local set function provides error checking whenever an object is created. CurrentValue The current value of the asset.

The following code block shows how the properties are defined. Note the set function defined for the Type property. It restricts the propertys values to one of three strings: bond, stock, or savings.
properties Description = ''; CurrentValue = 0; end properties(SetAccess = private) Date % Set value in constructor Type = 'savings'; % Provide a default value end
Constructor Method Code

The DocAsset class is not derived from another class, so you do not need to call a superclass constructor. MATLAB constructs an object when you assign values to the specified output argument (a in the following code):
function a = DocAsset(description,type,current_value) % DocAsset constructor if nargin > 0
18-5
18
a.Description = description; a.Date = date; a.Type = type; a.CurrentValue = current_value; end end % DocAsset
Set Function for Type Property

In this class design, there are only three types of assetsbonds, stocks, and savings. Therefore, the possible values for the Type property are restricted to one of three possible stings by defining a set function as follows:
function obj = set.Type(obj,type) if ~(strcmpi(type,'bond') || strcmpi(type,'stock') || strcmpi(type,'savings')) error('Type must be either bond, stock, or savings') end obj.Type = type; end %Type set function
The MATLAB runtime calls this function whenever an attempt is made to set the Type property, even from within the class constructor function or by assigning an initial value. Therefore, the following statement in the class definition would produce an error:
properties Type = 'cash'; end
The only exception is the set.Type function itself, where the statement:
obj.Type = type;
does not result in a recursive call to set.Type.
The DocAsset Display Method

The asset disp method is designed to be called from child-class disp methods. Its purpose is to display the data it stores for the child object. The method
18-6
simply formats the data for display in a way that is consistent with the formatting of the childs disp method:
function disp(a) % Display a DocAsset object fprintf('Description: %s\nDate: %s\nType: %s\nCurrentValue:%9.2f\n',... a.Description,a.Date,a.Type,a.CurrentValue); end % disp
The DocAsset subclass display methods can now call this method to display the data stored in the parent class. This approach isolates the subclass disp methods from changes to the DocAsset class.
Designing a Class for Stock Assets

Stocks are one type of asset. A class designed to store and manipulate information about stock holdings needs to contain the following information about the stock: The number of shares The price per share In addition, the base class (DocAsset) maintains general information including a description of the particular asset, the date the record was created, the type of asset, and its current value.

Open the DocStock class definition file in the MATLAB Editor. To use the class, create a folder named @DocStock and save DocStock.m to this folder. The parent folder of @DocStock must be on the MATLAB path.
Summary of the DocStock Class

This class is defined in one file, DocStock.m, which you must place in an @ folder of the same name. The parent folder of the @DocStock folder must be on the MATLAB path. See the addpath function for more information.
DocStock is a subclass of the DocAsset class.
18-7
18
The following table summarizes the properties defined for the DocStock class. DocStock Class Properties Name
NumShares SharePrice
Class
double double
Default
0 0
Description Number of shares of a particular stock Current value of asset Description of asset Current value of asset Date when record is created (set by date function) Type of asset (stock, bond, savings)
Properties Inherited from the DocAsset Class

Description CurrentValue Date Type char double char char '' 0 date ''
The following table summarizes the methods for the DocStock class. DocStock Class Methods Name
DocStock disp
Description Class constructor Displays information about the object
Specifying the Base Class

The < symbol specifies the DocAsset class as the base class for the DocStock class in the classdef line:
classdef DocStock < DocAsset
18-8

The following code shows how the properties are defined:
properties NumShares = 0; SharePrice = 0; end
Using the DocStock Class

Suppose you want to create a record of a stock asset for 200 shares of a company called Xdotcom with a share price of $23.47. Call the DocStock constructor function with the following arguments: Stock name or description Number of shares Share price For example, the following statement:
XdotcomStock = DocStock('Xdotcom',200,23.47);
creates a DocStock object, XdotcomStock, that contains information about a stock asset in Xdotcom Corp. The asset consists of 200 shares that have a per share value of $23.47.
The DocStock Constructor Method

The constructor first creates an instance of a DocAsset object since the DocStock class is derived from the DocAsset class (see The DocAsset Constructor Method on page 18-5). The constructor returns the DocStock object after setting value for its two properties:
function s = DocStock(description,num_shares,share_price) if nargin ~= 3 % Support no argument constructor syntax description = ''; num_shares = 0; share_price = 0; end
18-9
18
s = s@DocAsset(description,'stock',share_price*num_shares); s.NumShares = num_shares; s.SharePrice = share_price; end % DocStock
The DocStock disp Method

When you issue the statement (without terminating with a semicolon):
XdotcomStock = DocStock('Xdotcom',100,25)
the MATLAB runtime looks for a method in the @DocStock folder called disp. The disp method for the DocStock class produces this output:
Description: Xdotcom Date: 17-Nov-1998 Type: stock Current Value: $2500.00 Number of shares: 100 Share price: $25.00
The following function is the DocStock disp method. When this function returns from the call to the DocAsset disp method, it uses fprintf to display the Numshares and SharePrice property values on the screen:
function disp(s) disp@DocAsset(s) fprintf('Number of shares: %g\nShare price: %3.2f\n',... s.NumShares,s.SharePrice); end % disp
Designing a Class for Bond Assets

The DocBond class is similar to the DocStock class in that it is derived from the DocAsset class to represent a specific type of asset.

Open the DocBond class definition file in the MATLAB Editor.
18-10
To use the class, create a folder named @DocBond and save DocBond.m to this folder . The parent folder of @DocBond must be on the MATLAB path. See the addpath function for more information.
Summary of the DocBond Class

This class is defined in one file, DocBond.m, which you must place in an @ folder of the same name. The parent folder of the @DocBond folder must on the MATLAB path.
DocStock is a subclass of the DocAsset class.
The following table summarize the properties defined for the DocBond class DocBond Class Properties Name
FaceValue SharePrice
Class
double double
Default
0 0
Description Face value of the bond Current value of asset Description of asset Current value of asset Date when record is created (set by date function) Type of asset (stock, bond, savings)

Description CurrentValue Date Type char double char char '' 0 date ''
The following table summarizes the methods for the DocStock class.
18-11
18
DocBond Class Methods Name

DocBond disp calc_value
Description Class constructor Displays information about this object and calls the DocAsset disp method Utility function to calculate the bonds current value

The < symbol specifies the DocAsset class as the base class for the DocBond class in the classdef line:
classdef DocBond < DocAsset

The following code block shows how the properties are defined:
properties FaceValue = 0; Yield = 0; CurrentBondYield = 0; end
Using the DocBond Class

Suppose you want to create a record of an asset that consists of an xyzbond with a face value of $100 and a current yield of 4.3%. The current yield for the equivalent bonds today is 6.2%, which means that the market value of this particular bond is less than its face value. Call the DocBond constructor function with the following arguments: Bond name or description Bonds face value
18-12
Bonds interest rate or yield Current interest rate being paid by equivalent bonds (used to calculate the current value of the asset) For example, this statement:
b = DocBond('xyzbond',100,4.3,6.2);
creates a DocBond object, b, that contains information about a bond asset xyzbond with a face value of $100, a yield of 4.3%, and also contains information about the current yield of such bonds (6.2% in this case) that is used to calculate the current value. Note The calculations performed in this example are intended only to illustrate the use of MATLAB classes and do not represent a way to determine the actual value of any monetary investment.
The DocBond Constructor Method

The DocBond constructor method requires four arguments. It also supports the no argument syntax by defining default values for the missing input arguments:
function b = DocBond(description,face_value,yield,current_yield) if nargin ~= 4 description = ''; face_value = 0; yield = 0; current_yield = 0; end market_value = DocBond.calc_value(face_value,yield,current_yield); b = b@DocAsset(description,'bond',market_value); b.FaceValue = face_value; b.Yield = yield; b.CurrentBondYield = current_yield; end % DocBond
18-13
18
The calc_value Method

The DocBond class determines the market value of bond assets using a simple formula that scales the face value by the ratio of the bonds interest yield to the current yield for equivalent bonds. Calculation of the assets market value requires that the yields be nonzero, and should be positive just to make sense. While the calc_value method issues no errors for bad yield values, it does ensure bad values are not used in the calculation of market value. The assets market value is passed to the DocAsset base-class constructor when it is called within the DocBond constructor. calc_value has its Static attribute set to true because it does not accept a DocBond object as an input argument. The output of calc_value is used by the base-class (DocAsset) constructor:
methods (Static) function market_value = calc_value(face_value,yield,current_yield) if current_yield <= 0 || yield <= 0 market_value = face_value; else market_value = face_value*yield/current_yield; end end % calc_value end % methods
The DocBond disp Method

When you issue this statement (without terminating it with a semicolon):
b = DocBond('xyzbond',100,4.3,6.2)
the MATLAB runtime looks for a method in the @DocBond folder called disp. The disp method for the DocBond class produces this output:
Description: xyzbond Date: 17-Nov-1998 Type: bond Current Value: $69.35 Face value of bonds: $100 Yield: 4.30%
18-14
The following function is the DocBond disp method. When this function returns from the call to the DocAsset disp method, it uses fprintf to display the FaceValue, Yield, and CurrentValue property values on the screen:
function disp(b) disp@DocAsset(b) % Call DocAsset disp method fprintf('Face value of bonds: $%g\nYield: %3.2f%%\n',... b.FaceValue,b.Yield); end % disp
Designing a Class for Savings Assets

The DocSavings class is similar to the DocStock and DocBond class in that it is derived from the DocAsset class to represent a specific type of asset.

Open the DocSavings class definition file in the MATLAB Editor. To use the class, create a folder named @DocSavings and save DocSavings.m to this folder . The parent folder of @DocSavings must be on the MATLAB path.
Summary of the DocSavings Class

This class is defined in one file, DocSavings.m, which you must place in an @ folder of the same name. The parent folder of the @DocSavings folder must on the MATLAB path. See the addpath function for more information. The following table summarizes the properties defined for the DocSavings class. DocSavings Class Properties Name
InterestRate
Class
double
Default
''
Description Current interest rate paid on the savings account
18-15
18
DocSavings Class Properties (Continued) Name

Description CurrentValue Date
Class
char double char
Default
'' 0 date
Description Description of asset Current value of asset Date when record is created (set by date function) The type of asset (stock, bond, savings)
Type
char
''
The following table summarizes the methods for the DocSavings class. DocSavings Class Methods Name
DocSavings disp
Description Class constructor Displays information about this object and calls the DocAsset disp method

The < symbol specifies the DocAsset class as the base class for the DocBond class in the classdef line:
classdef DocSavings < DocAsset

The following code shows how the property is defined:
properties
18-16
InterestRate = 0; end
Using the DocSavings Class

Suppose you want to create a record of an asset that consists of a savings account with a current balance of $1000 and an interest rate of 2.9%. Call the DocSavings constructor function with the following arguments: Bank account description Account balance Interest rate paid on savings account For example, this statement:
sv = DocSavings('MyBank',1000,2.9);
creates a DocSavings object, sv, that contains information about an account in MyBank with a balance of $1000 and an interest rate of 2.9%.
The DocSavings Constructor Method

The savings account interest rate is saved in the DocSavings class InterestRate property. The asset description and the current value (account balance) are saved in the inherited DocAsset object properties. The constructor calls the base class constructor (DocAsset.m) to create an instance of the object. It then assigns a value to the InterestRate property. The constructor supports the no argument syntax by providing default values for the missing arguments.
function s = DocSavings(description,balance,interest_rate) if nargin ~= 3 description = ''; balance = 0; interest_rate = 0; end s = s@DocAsset(description,'savings',balance); s.InterestRate = interest_rate;
18-17
18
end % DocSavings
The DocSavings disp Method

When you issue this statement (without terminating it with a semicolon):
sv = DocSavings('MyBank',1000,2.9)
the MATLAB runtime looks for a method in the @DocSavings folder called disp. The disp method for the DocSavings class produces this output:
Description: MyBank Date: 17-Nov-1998 Type: savings Current Value: $1000.00 Interest Rate: 2.90%
The following function is the DocSaving disp method. When this function returns from the call to the DocAsset disp method, it uses fprintf to display the Numshares and SharePrice property values on the screen:
function disp(b) disp@DocAsset(b) % Call DocAsset disp method fprintf('%s%3.2f%%\n','Interest Rate: ',s.InterestRate); end % disp
18-18
Containing Assets in a Portfolio

Kinds of Containment
Aggregation is the containment of objects by other objects. The basic relationship is that each contained object "is a part of" the container object. Composition is a more strict form of aggregation in which the contained objects are parts of the containing object and are not associated with any other objects. Portfolio objects form a composition with asset objects because the asset objects are value classes, which are copied when the constructor method creates the DocPortfolio object. For example, consider a financial portfolio class as a container for a set of assets (stocks, bonds, savings, and so on). It can group, analyze, and return useful information about the individual assets. The contained objects are not accessible directly, but only via the portfolio class methods. A Simple Class Hierarchy on page 18-2 provides information about the assets collected by this portfolio class.
Designing the DocPortfolio Class

The DocPortfolio class is designed to contain the various assets owned by an individual client and to provide information about the status of his or her investment portfolio. This example implements a somewhat over-simplified portfolio class that: Contains an individuals assets Displays information about the portfolio contents Displays a 3-D pie chart showing the relative mix of asset types in the portfolio

Open the DocPortfolio class definition file in the MATLAB Editor. To use the class, create a folder named @DocPortfolio and save DocPortfolio.m to this folder . The parent folder of @DocPortfolio must be on the MATLAB path.
18-19
18
Summary of the DocPortfolio Class

This class is defined in one file, DocPortfolio.m, which you must place in an @ folder of the same name. The parent folder of the @DocPortfolio folder must on the MATLAB path. See the addpath function for more information. The following table summarizes the properties defined for the DocPortfolio class. DocPortfolio Class Properties Name
Name IndAssets TotalValue
Class
char cell double
Default
'' {} 0
Description Name of client owning the portfolio A cell array containing individual asset objects Value of all assets (calculated in the constructor method)
The following table summarizes the methods for the DocPortfolio class. DocBond Class Methods Name
DocPortfolio disp
Description Class constructor Displays information about this object and calls the DocAsset disp method Overloaded version of pie3 function designed to take a single portfolio object as an argument
pie3

The following code block shows how the properties are defined:
properties Name = ''; end
18-20
properties (SetAccess = private) IndAssets = {}; TotalValue = 0; end
How Class Properties Are Used

Name Stores the name of the client as a character string. The clients name is passed to the constructor as an input argument. IndAsset A cell array that stores asset objects (i.e., DocStock, DocBond, and DocSavings objects). These asset objects are passed to the DocPortfolio constructor as input arguments and assigned to the property from within the constructor function. IndAsset The structure of this property is known only to DocPortfolio class member functions so the propertys SetAccess attribute is set to private. TotalValue Stores the total value of the clients assets. The class constructor determines the value of each asset by querying the assets CurrentValue property and summing the result. Access to the TotalValue property is restricted to DocPortfolio class member functions by setting the propertys SetAccess attribute to private.
Using the DocPortfolio Class

The DocPortfolio class is designed to provide information about the financial assets owned by a client. There are three possible types of assets that a client can own: stocks, bonds, and savings accounts. The first step is to create an asset object to represent each type of asset owned by the client:
XYZStock = DocStock('XYZ Stocks',200,12.34); USTBonds = DocBond('U.S. Treasury Bonds',1600,3.2,2.8); SaveAccount = DocSavings('MyBank Acc # 123',2000,6); VictoriaSelna = DocPortfolio('Victoria Selna',... XYZStock,... SaveAccount,... USTBonds)
18-21
18
The DocPortfolio object displays the following information:

VictoriaSelna = Assets for Client: Victoria Selna Description: XYZ Stocks Date: 11-Mar-2008 Type: stock Current Value: $2468.00 Number of shares: 200 Share price: $12.34 Description: MyBank Acc # 123 Date: 11-Mar-2008 Type: savings Current Value: $2000.00 Interest Rate: 6.00% Description: U.S. Treasury Bonds Date: 11-Mar-2008 Type: bond Current Value: $1828.57 Face value of bonds: $1600 Yield: 3.20% Total Value: $6296.57
The DocPortfolio pie3 Method on page 18-23 provides a graphical display of the portfolio.
The DocPortfolio Constructor Method

The DocPortfolio constructor method takes as input arguments a clients name and a variable length list of asset objects (DocStock, DocBond, and DocSavings objects in this example). The IndAssets property is a cell array used to store all asset objects. From these objects, the constructor determines the total value of the clients assets. This value is stored in the TotalValue property:
function p = DocPortfolio(name,varargin)
18-22
if nargin > 0 p.Name = name; for k = 1:length(varargin) p.IndAssets{k} = varargin(k); asset_value = p.IndAssets{k}{1}.CurrentValue; p.TotalValue = p.TotalValue + asset_value; end end end % DocPortfolio
The DocPortfolio disp Method

The portfolio disp method lists the contents of each contained object by calling the objects disp method. It then lists the client name and total asset value:
function disp(p) fprintf('\nAssets for Client: %s\n',p.Name); for k = 1:length(p.IndAssets) disp(p.IndAssets{k}{1}) % Dispatch to corresponding disp end fprintf('\nTotal Value: $%0.2f\n',p.TotalValue); end % disp
The DocPortfolio pie3 Method

The DocPortfolio class overloads the MATLAB pie3 function to accept a portfolio object and display a 3-D pie chart illustrating the relative asset mix of the clients portfolio. MATLAB calls the @DocPortfolio/pie3.m version of pie3 whenever the input argument is a single portfolio object:
function pie3(p) % Step 1: Get the current value of each asset stock_amt = 0; bond_amt = 0; savings_amt = 0; for k = 1:length(p.IndAssets) if isa(p.IndAssets{k},'DocStock') stock_amt = stock_amt + p.IndAssets{k}.CurrentValue; elseif isa(p.IndAssets{k},'DocBond') bond_amt = bond_amt + p.IndAssets{k}.CurrentValue; elseif isa(p.IndAssets{k},'DocSavings') savings_amt = savings_amt + p.IndAssets{k}.CurrentValue; end % if
18-23
18
end % for % Step 2: Create labels and data for the pie graph k = 1; if stock_amt ~= 0 label(k) = {'Stocks'}; pie_vector(k) = stock_amt; k = k + 1; end % if if bond_amt ~= 0 label(k) = {'Bonds'}; pie_vector(k) = bond_amt; k = k + 1; end % if if savings_amt ~= 0 label(k) = {'Savings'}; pie_vector(k) = savings_amt; end % if % Step 3: Call pie3, adjust fonts and colors pie3(pie_vector,label);set(gcf,'Renderer','zbuffer') set(findobj(gca,'Type','Text'),... 'FontSize',14,'FontWeight','bold') colormap prism stg(1) = {['Portfolio Composition for ',p.Name]}; stg(2) = {['Total Value of Assets: $',num2str(p.TotalValue,'%0.2f')]}; title(stg,'FontSize',10) end % pie3
There are three parts in the overloaded pie3 method. Step 1 Get the CurrentValue property of each contained asset object and determine the total value in each category. Step 2 Create the pie chart labels and build a vector of graph data, depending on which objects are present in the portfolio. Step 3 Call the MATLAB pie3 function, make some font and colormap adjustments, and add a title.
18-24
Visualizing a Portfolio
You can use a DocPortfolio object to present an individuals financial portfolio. For example, given the following assets:
XYZStock = DocStock('XYZ Stocks',200,12.34); USTBonds = DocBond('U.S. Treasury Bonds',1600,3.2,2.8); SaveAccount = DocSavings('MyBank Acc # 123',2000,6); VictoriaSelna = DocPortfolio('Victoria Selna',... XYZStock,... SaveAccount,... USTBonds);
you can use the classs pie3 method to display the relative mix of assets as a pie chart.
pie3(VictoriaSelna)
18-25
18
18-26
Index
A
Index
arithmetic operators overloading 17-13
C
classes defining 3-5 value classes 5-4
objects as indices into objects 15-28 overloaded function 7-27 overloading 15-14 arithmetic operators 17-13 pie3 18-23
P
pie3 function overloaded 18-23
E
end method 15-27
polynomials example class 17-2
examples container class 18-19 polynomial class 17-2
R
reference, subscripted 15-14
M
methods
end 15-27
S
subscripted assignment 15-19 subscripting overloading 15-14 subsref 15-14
O
object-oriented programming overloading subscripting 15-14
V
value classes 5-4
Index-1

Matlab Oop

Uploaded by

Copyright:

Available Formats

Matlab Oop

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Matlab Oop

Uploaded by

Copyright:

Available Formats

MATLAB Object-Oriented Programming

How to Contact MathWorks

Web Newsgroup www.mathworks.com/contact_TS.html Technical Support

Using Object-Oriented Design in MATLAB

Begin Using Object-Oriented Programming

Video Demo of MATLAB Classes

MATLAB Programmer Without Object-Oriented Programming Experience

MATLAB Programmer with Object-Oriented Programming Experience

Begin Using Object-Oriented Programming

Comparing MATLAB with Other OO Languages on page 3-50

Using Object-Oriented Design in MATLAB

Why Use Object-Oriented Design

Approaches to Writing MATLAB Programs

Procedural Program Design

Object-Oriented Program Design

Why Use Object-Oriented Design

Classes and Objects

Using Objects in MATLAB Programs

Using Object-Oriented Design in MATLAB

Objects Organize Data

Why Use Object-Oriented Design

You could, for example, use a property reference in MATLAB functions:

Objects Manage Their Own Data

Using Object-Oriented Design in MATLAB

Why Use Object-Oriented Design

When Should You Start Creating Object-Oriented Programs

Understanding a Problem in Terms of Its Objects

Using Object-Oriented Design in MATLAB

Objects Manage Internal State

Why Use Object-Oriented Design

Using Object-Oriented Design in MATLAB

Defining Consistent Interfaces

Why Use Object-Oriented Design

n1 Properties Next Prev n2.Prev

n2 Properties Next Prev n2

n3 Properties Next Prev n2.Next

Using Object-Oriented Design in MATLAB

1 Unlink n2.Prev from n1 2 Unlink n1.Next from n2

n1 Properties Next Prev

n2 Properties Next Prev

n3 Properties Next Prev

n4 Properties Next Prev

Newly inserted node

Why Use Object-Oriented Design

Overloaded Functions and Operators

Reduce Code Redundancy

Using Object-Oriented Design in MATLAB

Class Diagram Notation

Class Diagram Notation

Using Object-Oriented Design in MATLAB

Properties AccountNumber AccountBalance

Employee Class Properties Name Address

FileReader (aggregation) has_a Car (composition)

What Is a Handle Object?

Using Object-Oriented Design in MATLAB

You can use gongSound2 to access the same audio source:

Play the gong sound with the new sample rate:

Modifying Handle Objects in Functions

Define a variable and pass it to myfunc:

The value of x has not changed after executing myFunc(x):