D52163GC10 sg2

Unauthorized reproduction or distribution prohibitedฺ Copyright© 2010, Oracle and/or its affiliatesฺ
n s e
i ce
b l el
fer a
a n s
o n -tr
Oracle Database n 11g: SQL Tuning
a
s eฺ
Workshop h a
e s ) u id
ฺ
ir ent G
ฺ m
d gp SVolume
tud
II • Student Guide
b i a@ this
i a (ja use
Ab to
o n io
n t
e A
J o s
D52163GC10
Edition 1.0
June 2008
D54948
Author Copyright © 2008, Oracle. All rights reserved.
Jean-François Verrier Disclaimer
This document contains proprietary information and is protected by copyright and

Technical Contributors other intellectual property laws. You may copy and print this document solely for your
own use in an Oracle training course. The document may not be modified or altered in
and Reviewers any way. Except where your use constitutes "fair use" under copyright law, you may
Muriel Fry (Special thanks) not use, share, download, upload, copy, print, display, perform, reproduce, publish,
license, post, transmit, or distribute this document in whole or in part without the
Joel Goodman express authorization of Oracle.
Harald van Breederode
The information contained in this document is subject to change without notice. If you
Jörn Bartels find any problems in the document, please report them in writing to: Oracle University,
500 Oracle Parkway, Redwood Shores, California 94065 USA. This document is not
Pekka Siltala warranted to be error-free.
Bernard Soleillant
Restricted Rights Notice
James Spiller If this documentation is delivered to the United States Government or anyone using
the documentation on behalf of the United States Government, the following notice is
n s e
Clay Fuller applicable:
ice
Ira Singer
b l el
Christopher Andrews
U.S. GOVERNMENT RIGHTS
fe r a
The U.S. Government’s rights to use, modify, reproduce, release, perform, display, or
Magnus Isaksson n s
disclose these training materials are restricted by the terms of the applicable Oracle
a
Sean Kim n - tr
license agreement and/or the applicable U.S. Government contract.
o
an
Trademark Notice
Trevor Bauwen
Yash Jain
s
a deฺ
Oracle is a registered trademark of Oracle Corporation and/or its affiliates. Other
h
)
s Gui
names may be trademarks of their respective owners.
e
r ฺ
Editors
p ฺ mi dent
Amitha Narayan d g Stu
Raj Kumar
b i a@ this
i a (ja use
b
Graphic Designer
A to
ni o
Priya Saxena
t o
e An
Jos
Publishers
Jothi Lakshmi
Veena Narasimhan
Contents
1 Exploring the Oracle Database Architecture

Objectives 1-2
Oracle Database Server Architecture: Overview 1-3
Connecting to the Database Instance 1-4
Oracle Database Memory Structures: Overview 1-6
Database Buffer Cache 1-7 n s e
i ce
Redo Log Buffer 1-8
b l el
Shared Pool 1-9
fe r a
Processing a DML Statement: Example 1-10
a n s
COMMIT Processing: Example 1-11
n - tr
o
Large Pool 1-12
s an
Java Pool and Streams Pool 1-13
) h a deฺ
Program Global Area (PGA) 1-14
r ฺ e s Gui
p ฺ mi dent
Background Process Roles 1-15
d g Stu
Automatic Shared Memory Management 1-16
i a@ this
Automated SQL Execution Memory Management 1-17
b
i a (ja use
Automatic Memory Management 1-18
b to
Database Storage Architecture 1-19
A
o n io
Logical and Physical Database Structures 1-21
nt
e A Segments, Extents, and Blocks 1-23
J o s SYSTEM and SYSAUX Tablespaces 1-24

Summary 1-25
Practice 1: Overview 1-26
2 Introduction to SQL Tuning

Objectives 2-2
Reasons for Inefficient SQL Performance 2-3
Inefficient SQL: Examples 2-4
Performance Monitoring Solutions 2-6
Monitoring and Tuning Tools: Overview 2-8
EM Performance Pages for Reactive Tuning 2-9
Tuning Tools: Overview 2-10
SQL Tuning Tasks: Overview 2-11
CPU and Wait Time Tuning Dimensions 2-12
Scalability with Application Design, Implementation, and Configuration 2-13
iii
Common Mistakes on Customer Systems 2-14
Proactive Tuning Methodology 2-16
Simplicity in Application Design 2-17
Data Modeling 2-18
Table Design 2-19

Index Design 2-20
Using Views 2-21
SQL Execution Efficiency 2-22
Writing SQL to Share Cursors 2-24
Performance Checklist 2-25
Summary 2-26
n s e
i ce
b l el
3 Introduction to the Optimizer
fer a
Objectives 3-2
a n s
Structured Query Language 3-3 n -tr
o
SQL Statement Representation 3-4
s an
SQL Statement Implementation 3-5
) h a deฺ
SQL Statement Processing: Overview 3-6
r ฺ e s Gui
ฺ
SQL Statement Processing: Steps 3-7
p mi dent
Step 1: Create a Cursor 3-8 d g Stu
i a@ this
Step 2: Parse the Statement 3-9
b
i a (ja use
Steps 3 and 4: Describe and Define 3-10
Ab to
Steps 5 and 6: Bind and Parallelize 3-11
n io
Steps 7 Through 9 3-12
o
nt
e A SQL Statement Processing PL/SQL: Example 3-13
J o s SQL Statement Parsing: Overview 3-14

Why Do You Need an Optimizer? 3-16
Optimization During Hard Parse Operation 3-18
Transformer: OR Expansion Example 3-19
Transformer: Subquery Unnesting Example 3-20
Transformer: View Merging Example 3-21
Transformer: Predicate Pushing Example 3-22
Transformer: Transitivity Example 3-23
Cost-Based Optimizer 3-24
Estimator: Selectivity 3-25
Estimator: Cardinality 3-26
Estimator: Cost 3-27
Plan Generator 3-28
Controlling the Behavior of the Optimizer 3-29
Optimizer Features and Oracle Database Releases 3-34
iv
Summary 3-35
4 Optimizer Operators
Objectives 4-2
Row Source Operations 4-3
Main Structures and Access Paths 4-4
Full Table Scan 4-5
Full Table Scans: Use Cases 4-6
ROWID Scan 4-7
Sample Table Scans 4-8
n s e
Indexes: Overview 4-10
i ce
Normal B*-tree Indexes 4-12
b l el
Index Scans 4-13
fer a
Index Unique Scan 4-14
a n s
Index Range Scan 4-15 n -tr
o
Index Range Scan: Descending 4-16
s an
Descending Index Range Scan 4-17
) h a deฺ
ฺ e s Gui
Index Range Scan: Function-Based 4-18
r
Index Full Scan 4-19
p ฺ mi dent
d g Stu
Index Fast Full Scan 4-20
i a@ this
Index Skip Scan 4-21
b
i a (ja use
Index Skip Scan: Example 4-23
Ab to
Index Join Scan 4-24
o n io
The AND-EQUAL Operation 4-25
nt
e A B*-tree Indexes and Nulls 4-26
J o s Using Indexes: Considering Nullable Columns 4-27
Index-Organized Tables 4-28
Index-Organized Table Scans 4-29
Bitmap Indexes 4-30
Bitmap Index Access: Examples 4-31
Combining Bitmap Indexes: Examples 4-32
Combining Bitmap Index Access Paths 4-33
Bitmap Operations 4-34
Bitmap Join Index 4-35
Composite Indexes 4-36
Invisible Index: Overview 4-37
Invisible Indexes: Examples 4-38
Guidelines for Managing Indexes 4-39
Investigating Index Usage 4-41
v
Clusters 4-44
When Are Clusters Useful? 4-45
Cluster Access Path: Examples 4-47
Sorting Operators 4-48
Buffer Sort Operator 4-49

Inlist Iterator 4-50
View Operator 4-51
Count Stop Key Operator 4-52
Min/Max and First Row Operators 4-53
Join Methods 4-54
Nested Loops Join 4-55
n s e
Nested Loops Join: Prefetching 4-56
i ce
Nested Loops Join: 11g Implementation 4-57
b l el
Sort Merge Join 4-58
fer a
Hash Join 4-59
a n s
Cartesian Join 4-60 n -tr
o
Join Types 4-61
s an
Equijoins and Nonequijoins 4-62
) h a deฺ
Outer Joins 4-63
r ฺ e s Gui
Semijoins 4-64
p ฺ mi dent
Antijoins 4-65 d g Stu
i a@ this
Other N-Array Operations 4-66
b
i a (ja use
Filter Operations 4-67
Ab to
Concatenation Operation 4-68
n io
UNION [ALL], INTERSECT, MINUS 4-69
o
nt
e A Result Cache Operator 4-70
J o s Summary 4-71
5 Interpreting Execution Plans

Objectives 5-2
What Is an Execution Plan? 5-3
Where to Find Execution Plans? 5-4
Viewing Execution Plans 5-6
The EXPLAIN PLAN Command 5-7
The EXPLAIN PLAN Command: Example 5-9
PLAN_TABLE 5-10
Displaying from PLAN_TABLE: Typical 5-11
Displaying from PLAN_TABLE: ALL 5-12
Displaying from PLAN_TABLE: ADVANCED 5-14
vi
AUTOTRACE 5-15
The AUTOTRACE Syntax 5-16
AUTOTRACE: Examples 5-17
AUTOTRACE: Statistics 5-18
Using the V$SQL_PLAN View 5-20

The V$SQL_PLAN Columns 5-21
The V$SQL_PLAN_STATISTICS View 5-22
Links Between Important Dynamic Performance Views 5-23
Querying V$SQL_PLAN 5-25
Automatic Workload Repository (AWR) 5-27
Managing AWR with PL/SQL 5-29
n s e
Important AWR Views 5-31 i ce
Querying the AWR 5-32 b l el
Generating SQL Reports from AWR Data 5-34 fer a
a n s
SQL Monitoring: Overview 5-35
n - tr
SQL Monitoring Report: Example 5-37 o
Interpreting an Execution Plan 5-40 s an
) h a deฺ
r ฺ e s Gui
Execution Plan Interpretation: Example 1 5-42
p ฺ mi dent
d g Stu
b i a@ this
Reading More Complex Execution Plans 5-48
i a (ja use
Reviewing the Execution Plan 5-49
Ab to
Looking Beyond Execution Plans 5-50
o n io
Summary 5-51
nt Practice 5: Overview 5-52
e A
J o s
6 Case Study: Star Transformation
Objectives 6-2
The Star Schema Model 6-3
The Snowflake Schema Model 6-4
Star Query: Example 6-5
Execution Plan Without Star Transformation 6-6
Star Transformation 6-7
Star Transformation: Considerations 6-8
Star Transformation: Rewrite Example 6-9
Retrieving Fact Rows from One Dimension 6-10
Retrieving Fact Rows from All Dimensions 6-11
Joining the Intermediate Result Set with Dimensions 6-12
Star Transformation Plan: Example 1 6-13
Star Transformation: Further Optimization 6-14
vii
Using Bitmap Join Indexes 6-15
Star Transformation Plan: Example 2 6-16
Star Transformation Hints 6-17
Bitmap Join Indexes: Join Model 1 6-18

Summary 6-22
7 Optimizer Statistics
n s e
Objectives 7-2
i ce
Optimizer Statistics 7-3
b l el
Types of Optimizer Statistics 7-4
fer a
Table Statistics (DBA_TAB_STATISTICS) 7-5
a n s
Index Statistics (DBA_IND_STATISTICS) 7-6 n -tr
o
Index Clustering Factor 7-8
s an
h
Column Statistics (DBA_TAB_COL_STATISTICS) 7-10
) a deฺ
Histograms 7-11 r ฺ e s Gui
Frequency Histograms 7-12 p ฺ mi dent
d g Stu
Viewing Frequency Histograms 7-13
b i a@ this
Height-Balanced Histograms 7-14
a (ja use
Viewing Height-Balanced Histograms 7-15
i
Ab to
Histogram Considerations 7-16
o n io
t
Multicolumn Statistics: Overview 7-17
n
e A Expression Statistics: Overview 7-19
J o s Gathering System Statistics 7-20
Gathering System Statistics: Example 7-22
Mechanisms for Gathering Statistics 7-24
Statistic Preferences: Overview 7-25
When to Gather Statistics Manually 7-27
Manual Statistics Gathering 7-28
Manual Statistics Collection: Factors 7-29
Managing Statistics Collection: Example 7-30
Optimizer Dynamic Sampling: Overview 7-31
Optimizer Dynamic Sampling at Work 7-32
OPTIMIZER_DYNAMIC_SAMPLING 7-33
Locking Statistics 7-34
Summary 7-35
viii
8 Using Bind Variables
Objectives 8-2
Cursor Sharing and Different Literal Values 8-3
Cursor Sharing and Bind Variables 8-4
Bind Variables in SQL*Plus 8-5

Bind Variables in Enterprise Manager 8-6
Bind Variable Peeking 8-7
Cursor Sharing Enhancements 8-9
The CURSOR_SHARING Parameter 8-11
Forcing Cursor Sharing: Example 8-12
Adaptive Cursor Sharing: Overview 8-13
n s e
Adaptive Cursor Sharing: Architecture 8-14
i ce
Adaptive Cursor Sharing: Views 8-16
b l el
Adaptive Cursor Sharing: Example 8-18
fer a
Interacting with Adaptive Cursor Sharing 8-19
a n s
Summary 8-20 n -tr
o
s an
) h a deฺ
9 Using Optimizer Hints
r ฺ e s Gui
Objectives 9-2
p ฺ mi dent
d
Optimizer Hints: Overview g9-3 Stu
i
Types of Hints b9-4 a@ this
Specifying
a ja 9-5use
(Hints
i
bfor Hints to9-6
o A
Rules
t o ni Hint Recommendations 9-7

An Optimizer Hint Syntax: Example 9-8
e
Jos Hint Categories 9-9
Optimization Goals and Approaches 9-10
Hints for Access Paths 9-11
The INDEX_COMBINE Hint: Example 9-15
Hints for Query Transformation 9-17
Hints for Join Orders 9-20
Hints for Join Operations 9-21
Additional Hints 9-23
Hints and Views 9-26
Global Table Hints 9-28
Specifying a Query Block in a Hint 9-29
Specifying a Full Set of Hints 9-30
Summary 9-31
ix
10 Application Tracing
Objectives 10-2
End-to-End Application Tracing Challenge 10-3
End-to-End Application Tracing 10-4
Location for Diagnostic Traces 10-5

What Is a Service? 10-6
Use Services with Client Applications 10-7
Tracing Services 10-8
Use Enterprise Manager to Trace Services 10-10
Service Tracing: Example 10-11
Session Level Tracing: Example 10-12
n s e
Trace Your Own Session 10-13
i ce
The trcsess Utility 10-14
b l el
Invoking the trcsess Utility 10-15
fer a
The trcsess Utility: Example 10-16 a n s
SQL Trace File Contents 10-17 n -tr
o
SQL Trace File Contents: Example 10-18
s an
Formatting SQL Trace Files: Overview 10-19
) h a deฺ
Invoking the tkprof Utility 10-20 r ฺ e s Gui
tkprof Sorting Options 10-22 p ฺ mi dent
Output of the tkprof Command d g10-24Stu
tkprof Output with a Nob ia@Example
Index: t h is 10-29
tkprof Outputia
(j
with Index:u se 10-30
Example
A b to
Summaryo
ni 10: Overview 10-32
10-31
t o
An
Practice
e
Jos 11 Automating SQL Tuning
Objectives 11-2
Tuning SQL Statements Automatically 11-3
Application Tuning Challenges 11-4
SQL Tuning Advisor: Overview 11-5
Stale or Missing Object Statistics 11-6
SQL Statement Profiling 11-7
Plan Tuning Flow and SQL Profile Creation 11-8
SQL Tuning Loop 11-9
Access Path Analysis 11-10
SQL Structure Analysis 11-11
SQL Tuning Advisor: Usage Model 11-12
Database Control and SQL Tuning Advisor 11-13
Running SQL Tuning Advisor: Example 11-14
x
Implementing Recommendations 11-15
SQL Access Advisor: Overview 11-16
SQL Access Advisor: Usage Model 11-17
Possible Recommendations 11-18
SQL Access Advisor Session: Initial Options 11-19

SQL Access Advisor: Workload Source 11-21
SQL Access Advisor: Recommendation Options 11-22
SQL Access Advisor: Schedule and Review 11-23
SQL Access Advisor: Results 11-24
SQL Access Advisor: Results and Implementation 11-25
SQL Tuning Loop 11-26
n s e
Automatic SQL Tuning 11-27
i ce
Automatic Tuning Process 11-28
b l el
Automatic SQL Tuning Controls 11-30
fer a
Automatic SQL Tuning Task 11-31
a n s
Configuring Automatic SQL Tuning 11-32 n - tr
o
Automatic SQL Tuning: Result Summary 11-33
s an
h
Automatic SQL Tuning: Result Details 11-34
) a deฺ
e s Gui
Automatic SQL Tuning Result Details: Drilldown 11-35
r ฺ
ฺ mi dent
Automatic SQL Tuning Considerations 11-36
p
Summary 11-37 d g Stu
i a@ this
b
i a (ja use
A
Appendix to and Solutions
b A: Practices
ni o
n t o
e A Index
Jos
xi
J
o s e
n A
t o n
Aio
b i a
b i
to
(ja use
d p ฺ
a@ this
xii
r ฺ
g Stu
e ) h
mi dent
s
s Gui
an
o
a deฺ
tr n -
a n s fe
r
a b l el
i
ce n s e
Appendix A
Practices and Solutions
n s e
i ce
b l el
fer a
a n s
n -tr
o
s an
) h a deฺ
r ฺ e s Gui
p ฺ mi dent
d g Stu
b i a@ this
i a (ja use
Ab to
o n io
nt
e A
J o s
Table of Contents
Practices for Lesson 1 ......................................................................................................... 3
Practice 1-1: Exploring the Database Architecture......................................................... 4
Practices for Lesson 2 ....................................................................................................... 12
Practice 2-1: Avoiding Common Mistakes................................................................... 13

Practice 3-1: Understanding Optimizer Decisions........................................................ 33
Practice 4-1: Using Different Access Paths .................................................................. 58
Practice 4-2: Using the Result Cache.......................................................................... 121
Practices for Lesson 5 ..................................................................................................... 144
Practice 5-1: Extracting Execution Plans.................................................................... 145 n s e
Practices for Lesson 6 ..................................................................................................... 179 ice
Practice 6-1: Star Schema Tuning............................................................................... 180 b l el
fe ra
n s
Practice 7-1: Using System Statistics ......................................................................... 197
a
- tr
Practice 7-2: Automatic Statistics Gathering.............................................................. 212
n
o
an
s
) h a deฺ
Practice 8-1: Understanding Adaptive Cusrsor Sharing ............................................. 241
r ฺ e s Gui
Practice 8-2: Understanding CURSOR_SHARING................................................... 260
p ฺ mi dent
Practice 9-1: Using Hints ............................................................................................ 274
d g Stu
Practices for Lesson 10 ................................................................................................... 318
i a@ this
Practice 10-1: Tracing Applications ........................................................................... 319
b
i a (ja use
Practices for Lesson 11 ................................................................................................... 363
b to
Practice 11-1: Proactively Tuning High-Load SQL Statements................................. 364
A
o n io
Practice 11-2: Using SQL Access Advisor................................................................. 385
ntPractice 11-3: Using Automatic SQL Tuning............................................................. 415
e A
J o s
Oracle Database 11g: SQL Tuning Workshop A - 2

Practices for Lesson 1
n s e
i ce
b l el
fer a
a n s
n -tr
o
s an
) h a deฺ
r ฺ e s Gui
p ฺ mi dent
d g Stu
b i a@ this
i a (ja use
Ab to
o n io
nt
e A
J o s

Practice 1-1: Exploring the Database Architecture
Fill in the blanks with the correct answers:
1) The two main components of Oracle RDBMS are
_________________________ and _______________________.
2) An instance consists of _____________________and _____________________

processes.
3) A session is a connection between the _______________ process and either the

______________ process or the ________________ process. e
n s
i ce
4) Name some components of the System Global Area (SGA). b l el
fer a
• ___________________________
a n s
• ___________________________ n - tr
o
• ___________________________ s an
) h a deฺ
• ___________________________
r ฺ e s Gui
• ___________________________
p ฺ mi dent
d g Stu
b i a@ this
• ___________________________
i a (ja use
• ____________________________
A b to
o
ni some background processes:
5)toList
An
e •
Jos
___________________________
• ___________________________
• ___________________________
• ___________________________
• ___________________________
• ___________________________
6) The _______________________ process writes the dirty buffers to the data files.
7) The _______________________ process writes the redo logs to the log files.
8) Name some files associated with an Oracle database.

Practice 1-1: Exploring the Database Architecture (continued)
• ___________________________
• ___________________________
• ___________________________
• ___________________________
• ___________________________
• ___________________________
• ____________________________
9) Some of the logical storage structures of an Oracle database are: e

n s
• ___________________________ i ce
b l el
• ___________________________
fer a
• ___________________________ a n s
• ___________________________ o n -tr
• ___________________________ s an
) h a deฺ
• ___________________________ ฺes
r G ui
ฺ
• ____________________________
p mi dent
d g Stu
b i a@ this
i a (ja useprocess copies the redo log files to an archive
10) The ___________________
b
destination.
A to
ni o
t o
e An11) The _____________________ contains data and control information for a server or a
Jos background process.
12) The logical tablespace structure is associated with the physical

____________________ files on disk.
13) State whether the following statements are true or false.

a) The SGA includes the Database buffer cache and the Redo log buffer. ____
b) Each server process and background process has its own Program Global Area
(PGA). ____
c) User processes run the application or Oracle tool code. ____
d) Oracle Database processes include server processes and background processes.
____.

14) From a terminal session connected as the oracle user, execute the
processes.sh script located in your
$HOME/solutions/Database_Architecture directory. What does this
script show you?
a) It shows you all the database instance processes currently running on your
machine. This includes both background processes and foreground processes.

[oracle@edrsr33p1-orcl Database_Architecture]$ ./processes.sh
oracle 9537 1 0 01:00 ? 00:00:00 ora_w000_orcl
oracle 12132 22002 0 Mar26 pts/4 00:01:08
/u01/app/oracle/product/11.1.0/db_1/jdk/bin/java -server -Xmx256M -
XX:MaxPermSize=200m -XX:MinHeapFreeRatio=20 -XX:MaxHeapFreeRatio=40
-DORACLE_HOME=/u01/app/oracle/product/11.1.0/db_1 -
Doracle.home=/u01/app/oracle/product/11.1.0/db_1/oc4j -
n s e
ce
Doracle.oc4j.localhome=/u01/app/oracle/product/11.1.0/db_1/edrsr33p1
.us.oracle.com_orcl/sysman -
eli
DEMSTATE=/u01/app/oracle/product/11.1.0/db_1/edrsr33p1.us.oracle.com
a b l
_orcl -Doracle.j2ee.dont.use.memory.archive=true -
fe r
Djava.protocol.handler.pkgs=HTTPClient -
a n s
- tr
Doracle.security.jazn.config=/u01/app/oracle/product/11.1.0/db_1/oc4
n
o
j/j2ee/OC4J_DBConsole_edrsr33p1.us.oracle.com_orcl/config/jazn.xml -
an
Djava.security.policy=/u01/app/oracle/product/11.1.0/db_1/oc4j/j2ee/
s
) h a deฺ
OC4J_DBConsole_edrsr33p1.us.oracle.com_orcl/config/java2.policy -
onfig/OCMTrustedCerts.txt-
r ฺ s Gui
Djavax.net.ssl.KeyStore=/u01/app/oracle/product/11.1.0/db_1/sysman/c
e
ฺ mi dent
Djava.security.properties=/u01/app/oracle/product/11.1.0/db_1/oc4j/j
p
g Stu
2ee/home/config/jazn.security.props -
d
a@ this
DEMDROOT=/u01/app/oracle/product/11.1.0/db_1/edrsr33p1.us.oracle.com
b i
_orcl -Dsysman.md5password=true -
i a (ja use
Drepapi.oracle.home=/u01/app/oracle/product/11.1.0/db_1 -
b to
Ddisable.checkForUpdate=true -
A
io
Doracle.sysman.ccr.ocmSDK.websvc.keystore=/u01/app/oracle/product/11
t o n
.1.0/db_1/jlib/emocmclnt.ks -
A n Dice.pilots.html4.ignoreNonGenericFonts=true -
o s e Djava.awt.headless=true -jar
J /u01/app/oracle/product/11.1.0/db_1/oc4j/j2ee/home/oc4j.jar -config
/u01/app/oracle/product/11.1.0/db_1/oc4j/j2ee/OC4J_DBConsole_edrsr33
p1.us.oracle.com_orcl/config/server.xml
oracle 12225 1 0 Mar26 ? 00:00:00 ora_pmon_orcl
oracle 12227 1 0 Mar26 ? 00:00:00 ora_vktm_orcl
oracle 12231 1 0 Mar26 ? 00:00:00 ora_diag_orcl
oracle 12233 1 0 Mar26 ? 00:00:00 ora_dbrm_orcl
oracle 12235 1 0 Mar26 ? 00:00:00 ora_psp0_orcl
oracle 12239 1 0 Mar26 ? 00:00:00 ora_dia0_orcl
oracle 12241 1 0 Mar26 ? 00:00:00 ora_mman_orcl
oracle 12243 1 0 Mar26 ? 00:00:21 ora_dbw0_orcl
oracle 12245 1 0 Mar26 ? 00:00:48 ora_lgwr_orcl
oracle 12247 1 0 Mar26 ? 00:00:01 ora_ckpt_orcl
oracle 12249 1 0 Mar26 ? 00:00:04 ora_smon_orcl
oracle 12251 1 0 Mar26 ? 00:00:00 ora_reco_orcl
oracle 12253 1 0 Mar26 ? 00:00:05 ora_mmon_orcl
oracle 12255 1 0 Mar26 ? 00:00:01 ora_mmnl_orcl
oracle 12257 1 0 Mar26 ? 00:00:00 ora_d000_orcl
oracle 12259 1 0 Mar26 ? 00:00:00 ora_s000_orcl
oracle 12282 1 0 Mar26 ? 00:00:00 ora_fbda_orcl

oracle 12284 1 0 Mar26 ? 00:00:00 ora_smco_orcl
oracle 12290 1 0 Mar26 ? 00:00:00 ora_qmnc_orcl
oracle 12302 1 0 Mar26 ? 00:00:03 oracleorcl
(LOCAL=NO)
oracle 12325 1 0 Mar26 ? 00:00:00 ora_q000_orcl
(LOCAL=NO)
(LOCAL=NO)
(LOCAL=NO)
(LOCAL=NO)
(LOCAL=NO)
n s e
(LOCAL=NO)
i ce
oracle 12362 1 0 Mar26 ? 00:00:00 ora_q001_orcl
b l el
oracle
oracle
12570
20119
1 0 Mar26 ?
1 0 14:50 ?
00:00:01 ora_cjq0_orcl
00:00:00 oracleorcl
fe r a
(LOCAL=NO)
a n s
oracle 20482 20480 0 14:56 pts/2 -tr
00:00:00 grep orcl
o n
an
oracle 22002 1 0 Mar26 pts/4 00:00:08
/u01/app/oracle/product/11.1.0/db_1/perl/bin/perl
h s
a deฺ
/u01/app/oracle/product/11.1.0/db_1/bin/emwd.pl dbconsole
)
s Gui
/u01/app/oracle/product/11.1.0/db_1/edrsr33p1.us.oracle.com_orcl/sys
e
r ฺ
mi dent
man/log/emdb.nohup
p ฺ
[oracle@edrsr33p1-orcl Database_Architecture]$
g Stu
d
b i a@ this
--------------------------------------------------------------
#!/bin/bash (ja se
b i a to u
i o
ps -efA| grep orcl
nton
e A 15) From a terminal session connected as the oracle user, execute the files.sh
J o s script located in your $HOME/solutions/Database_Architecture
directory. What does this script show you?
a) This script shows you the location and names of all database files, the
initialization file, the password file, and trace files.
[oracle@edrsr33p1-orcl Database_Architecture]$ ./files.sh
SQL*Plus: Release 11.1.0.6.0 - Production on Thu Mar 27 17:58:56

2008
Copyright (c) 1982, 2007, Oracle. All rights reserved.
Connected to:
Oracle Database 11g Enterprise Edition Release 11.1.0.6.0 -
Production
With the Partitioning, Oracle Label Security, OLAP, Data Mining
and Real Application Testing options

SQL>
SQL> col name format a45
SQL>
SQL> select name from v$controlfile;
NAME
---------------------------------------------
/u01/app/oracle/oradata/orcl/control01.ctl
SQL>
SQL>
SQL> col member format a45
SQL>
SQL> select group#,member from v$logfile;
n s e
i ce
GROUP# MEMBER
b l el
----------
3
---------------------------------------------
/u01/app/oracle/oradata/orcl/redo03.log
fer a
2 /u01/app/oracle/oradata/orcl/redo02.log
a n s
1 /u01/app/oracle/oradata/orcl/redo01.log
o n -tr
SQL>
s an
SQL> col tablespace_name format a20 ) h
SQL> a deฺ
SQL> col file_name format a45 rฺe
s Gui
SQL> ฺ m i ent
d p tud from dba_data_files;
g file_name
SQL> select tablespace_name,
@ his S
TABLESPACE_NAME bia FILE_NAME
t
( j a
-------------------- s e
---------------------------------------------
USERS bia
t o u /u01/app/oracle/oradata/orcl/users01.dbf
i oA
UNDOTBS1
n
/u01/app/oracle/oradata/orcl/undotbs01.dbf
ntoSYSTEM
SYSAUX /u01/app/oracle/oradata/orcl/sysaux01.dbf
e A /u01/app/oracle/oradata/orcl/system01.dbf
J o s EXAMPLE /u01/app/oracle/oradata/orcl/example01.dbf
SQL>
SQL> select tablespace_name, file_name from dba_temp_files;
TABLESPACE_NAME FILE_NAME
-------------------- ---------------------------------------------
TEMP /u01/app/oracle/oradata/orcl/temp01.dbf
SQL>
SQL> exit;
Disconnected from Oracle Database 11g Enterprise Edition Release
11.1.0.6.0 - Production
-rw-rw---- 1 oracle oinstall 1544 Aug 22 2007
/u01/app/oracle/product/11.1.0/db_1/dbs/hc_orcl.dat
-rw-r----- 1 oracle oinstall 1536 Mar 26 22:03
/u01/app/oracle/product/11.1.0/db_1/dbs/orapworcl
-rw-r----- 1 oracle oinstall 2560 Mar 27 03:13
/u01/app/oracle/product/11.1.0/db_1/dbs/spfileorcl.ora

alert cdump hm incident incpkg ir lck metadata stage sweep
trace
-rw-r--r-- 1 oracle oinstall 557386 Mar 27 13:00
/u01/app/oracle/diag/rdbms/orcl/orcl/trace/alert_orcl.log
--------------------------------------------------------------
#!/bin/bash
cd /home/oracle/solutions/Database_Architecture
export ORACLE_SID=orcl
export ORACLE_HOME=/u01/app/oracle/product/11.1.0/db_1
n s e
export
i ce
PATH=/u01/app/oracle/product/11.1.0/db_1/bin:/bin:/usr/bin:/usr/loca
b l el
l/bin:/usr/X11R6/bin:/usr/java/jdk1.5.0_11/bin:/bin
fe r a
a n s
sqlplus / as sysdba @files.sql
o n -tr
ls -l $ORACLE_HOME/dbs/*orcl*
s an
) h a deฺ
ls /u01/app/oracle/diag/rdbms/orcl/orcl
r ฺ e s Gui
ฺ mi dent
ls -l /u01/app/oracle/diag/rdbms/orcl/orcl/trace/alert*
p
d g Stu
b i a@ this
--------------------------------------------------------------
set echo on (ja se

i a o
b format ta45 u
o A
ni name from v$controlfile;
col name
ntoselect
e A
J o s
col member format a45
select group#,member from v$logfile;
col tablespace_name format a20

col file_name format a45
select tablespace_name, file_name from dba_data_files;
select tablespace_name, file_name from dba_temp_files;
exit;
16) From a terminal session connected as the oracle user, execute the sga.sh script
located in your $HOME/solutions/Database_Architecture directory.
What does this script show you?

a) This script prints the various pools held in your SGA.
[oracle@edrsr33p1-orcl Database_Architecture]$ ./sga.sh

2008
Connected to:
Production
n s e
SQL>
i ce
SQL> select * from v$sgainfo;
b l el
NAME BYTES RES
fer a
-------------------------------- ---------- ---
a n s
Fixed SGA Size 1303188 No
n - tr
Redo Buffers 5181440 No o
Buffer Cache Size 335544320 Yes
s an
Shared Pool Size a deฺ
276824064 Yes
) h
Large Pool Size
r ฺ e s Gui
4194304 Yes
mi dent
Java Pool Size 12582912 Yes
Streams Pool Size
p ฺ
g Stu
0 Yes
Shared IO Pool Size
d 0 Yes
a@ this
Granule Size 4194304 No
Maximum SGA Size
b i 845348864 No
i a (ja use
Startup overhead in Shared Pool 46137344 No
Ab to
o n io
NAME BYTES RES
-------------------------------- ---------- ---
nt
e A Free SGA Memory Available 209715200
J o s 12 rows selected.
SQL>
SQL> col component format a30
SQL>
SQL> select component,current_size,min_size,max_size from
v$sga_dynamic_components;
COMPONENT CURRENT_SIZE MIN_SIZE MAX_SIZE

------------------------------ ------------ ---------- ----------
shared pool 276824064 226492416 276824064
large pool 4194304 4194304 4194304
java pool 12582912 12582912 12582912
streams pool 0 0 0
DEFAULT buffer cache 335544320 335544320 385875968
KEEP buffer cache 0 0 0
RECYCLE buffer cache 0 0 0
DEFAULT 2K buffer cache 0 0 0

COMPONENT CURRENT_SIZE MIN_SIZE MAX_SIZE

------------------------------ ------------ ---------- ----------
Shared IO Pool 0 0 0
ASM Buffer Cache 0 0 0
14 rows selected.
SQL>
SQL> exit;
n s e
i ce
b l el
--------------------------------------------------------------
fer a
a n s
#!/bin/bash
o n -tr
cd /home/oracle/solutions/Database_Architecture
s an
) h a deฺ
r ฺ e s Gui
ฺ mi dent
p
d g Stu
export
b i a@ this
i a (ja use
A b to
ni o
ntosqlplus / as sysdba @sga.sql
e A
J o s --------------------------------------------------------------
set echo on
select * from v$sgainfo;
col component format a30
select component,current_size,min_size,max_size from

v$sga_dynamic_components;
exit;

n s e
i ce
b l el
fer a
a n s
n -tr
o
s an
) h a deฺ
r ฺ e s Gui
p ฺ mi dent
d g Stu
b i a@ this
i a (ja use
Ab to
o n io
nt
e A
J o s

Practice 2-1: Avoiding Common Mistakes
In this practice, you examine some common mistakes in writing SQL statements. You
have to find a workaround to enhance performance.
1) You analyze a correlated subquery first. Before executing the culprit statement,
execute the correlation_setup.sh script to set up the environment for this
example. Make sure you run the script from a terminal session connected as the
oracle user. You can find the scripts for all the following cases in your
$HOME/solutions/Common_Mistakes directory.
[oracle@edrsr33p1-orcl Common_Mistakes]$ ./correlation_setup.sh
SQL*Plus: Release 11.1.0.6.0 - Production on Wed Mar 26 20:21:50

2008
n s e
i ce
el
a b l
fe r
Connected to:
a n s
Production
o n -tr
an
s
) h a deฺ
SQL> SQL>
r ฺ e s Gui
User altered.
p ฺ mi dent
SQL> SQL> d g Stu
Grant succeeded. ia@ is
b t h
SQL> SQL>ia (ja usefrom Oracle Database 11g Enterprise Edition
Disconnected
A
Release to - Production
b11.1.0.6.0
o
ni Real
ntoand Application Testing options
e A [oracle@edrsr33p1-orcl Common_Mistakes]$
J o s
--------------------------------------------------------------
#!/bin/bash
cd /home/oracle/solutions/Common_Mistakes
export
sqlplus / as sysdba <<EOF
alter user sh identified by sh account unlock;
grant dba to sh;

EOF

Practice 2-1: Avoiding Common Mistakes (continued)
2) Connected as the SH user from a SQL*Plus session (stay connected to that session
until the end of this case), make sure you execute the following command:
set timing on
@flush
The goal of the first command is to tell you how long the next command takes to
execute. The flush.sql script flushes both the shared pool and the buffer cache to
avoid most caching effects so that you can have good comparisons between two
executions. Note: You should not use commands found in this script on a production
system.
[oracle@edrsr33p1-orcl Common_Mistakes]$ sqlplus sh/sh

2008
n s e
i ce
b l el
fe r a
Connected to:
a n s
-
n tr
Production
o
an
s
) h a deฺ
SQL> set timing on
r ฺ e s Gui
SQL> @flush
p ฺ mi dent
d g Stu
a@ this
System altered.
b i
i a (ja use
Elapsed: 00:00:00.21
A b to
ni o
System altered.
ntoElapsed: 00:00:00.49
e A SQL>
J o s
--------------------------------------------------------------
alter system flush shared_pool;

alter system flush buffer_cache;
3) From the same SQL*Plus session, execute the following statement and note the time
it takes to execute (You can use the correlation.sql script.):
SELECT COUNT(*)
FROM products p
WHERE prod_list_price < 1.15 * (SELECT avg(unit_cost)
FROM costs c
WHERE c.prod_id = p.prod_id);
SQL> @correlation
SQL>
SQL> SELECT COUNT(*)
2 FROM products p
3 WHERE prod_list_price < 1.15 * (SELECT avg(unit_cost)

4 FROM costs c
5 WHERE c.prod_id = p.prod_id);
COUNT(*)
----------
46
Elapsed: 00:00:01.21
SQL>
--------------------------------------------------------------
set timing on
set echo on
SELECT COUNT(*)
n s e
FROM products p
i ce
WHERE prod_list_price < 1.15 * (SELECT avg(unit_cost)
b l el
FROM costs c
WHERE c.prod_id = p.prod_id);
fe r a
a n s
n - tr
o
4) Before trying to fix the previous statement, flush your environment again using the
flush.sql script from the SQL*Plus session.
s an
) h a deฺ
SQL> @flush
r ฺ e s Gui
ฺ m i ent
SQL> alter system flush shared_pool;
System altered.
d p
g Stud
b i
Elapsed: 00:00:00.10 a@ this
(ja flush
SQL> alter system
i a u se buffer_cache;
A b to
i o
System altered.
nton
e A Elapsed: 00:00:00.17
J o s SQL>
5) How do you rewrite this statement to enhance performance? Test your solution. You
should discuss this with your instructor.
SQL> @nocorrelation
SQL>
SQL> set timing on
SQL> set echo on
SQL>
SQL> SELECT COUNT(*)
2 FROM products p, (SELECT prod_id, AVG(unit_cost) ac FROM
costs GROUP BY prod_id) c
3 WHERE p.prod_id = c.prod_id AND
4 p.prod_list_price < 1.15 * c.ac;
COUNT(*)
----------
46
Elapsed: 00:00:00.16

SQL>
--------------------------------------------------------------
set timing on
set echo on
SELECT COUNT(*)
FROM products p, (SELECT prod_id, AVG(unit_cost) ac FROM costs
GROUP BY prod_id) c
WHERE p.prod_id = c.prod_id AND
p.prod_list_price < 1.15 * c.ac;
6) Exit from your SQL*Plus session.

n s e
ce
SQL> exit
eli
a b l
fe r
a n s
[oracle@edrsr33p1-orcl Common_Mistakes]$
o n -tr
s an
7) From your terminal session, execute the correlation_cleanup.sh script to
clean up your environment.
) h a deฺ
r ฺ e G ui
s ./correlation_cleanup.sh
ฺ m i ent
SQL*Plus: Release 11.1.0.6.0
d p
g Stud - Production on Wed Mar 26 20:49:22
ia@
2008
b t h is
Copyright (c)ja e
( 1982,
u s
2007, Oracle. All rights reserved.
A bia to
o n io
Connected to:
t
An Production
e
Jos With the Partitioning, Oracle Label Security, OLAP, Data Mining
SQL> SQL>
Revoke succeeded.
SQL> SQL> Disconnected from Oracle Database 11g Enterprise Edition

Release 11.1.0.6.0 - Production
--------------------------------------------------------------
#!/bin/bash

export
revoke dba from sh;
EOF
8) Before continuing, execute the setup_rest.sh script to set up the environment

for all the examples that follow. Make sure you run the script from a terminal session
connected as the oracle user. You can find the scripts for all the following cases in n s e
ice
your $HOME/solutions/Common_Mistakes directory.
b l el
[oracle@edrsr33p1-orcl Common_Mistakes]$ ./setup_rest.sh
f er a
n s
a20:59:53
SQL*Plus: Release 11.1.0.6.0 - Production on Wed Mar 26
- t r
2008
n on
s a reserved.
a eฺ
Copyright (c) 1982, 2007, Oracle. All rights
) h i d
i r ฺes nt Gu
Connected to:
g p
Oracle Database 11g Enterprise ฺm Edition
u d e Release 11.1.0.6.0 -
Production d S t
i a @ Oracle
h i s
With the Partitioning,
b t Label Security, OLAP, Data Mining
i a (ja use
A
SQL> SQL> to drop user jfv cascade
b SQL> SQL>
o
ni at line *
ntoERROR 1:
e A ORA-01918: user 'JFV' does not exist
J o s
SQL> SQL>
User created.
SQL> SQL>
Grant succeeded.
SQL> SQL> Connected.

SQL> SQL> drop table orders purge
*
ERROR at line 1:
ORA-00942: table or view does not exist
SQL> SQL>
Table created.
SQL> SQL> 2 3 4 5 6 7
PL/SQL procedure successfully completed.

SQL> SQL> 2 3 4 5 6 7
SQL> SQL> drop table employees purge

*
ERROR at line 1:
SQL> drop table job_history purge

*
ERROR at line 1:
n s e
SQL> SQL>
i ce
Table created.
b l el
SQL> SQL> 2 3 4 5 6 7
fer a
a n s
n -tr
o
an
SQL> SQL>
Table created.
s
a deฺ
)7h
SQL> SQL> 2 3 4 5
r
6
ฺ e s Gui
ฺ m i ent
SQL> SQL>
d g p t u d
Index created.
a@ this S
b i
SQL> SQL> SQL>
i a (jadroputable
se old purge
b
at line 1:t
ERROR A
* o
ni o
ntoORA-00942: table or view does not exist
e A
J o s SQL> drop table new purge
*
ERROR at line 1:
SQL> SQL>
Table created.
SQL>
Table created.
SQL> SQL> 2 3 4 5 6 7
SQL> SQL> 2 3 4 5 6 7
SQL> SQL> SQL> Disconnected from Oracle Database 11g Enterprise

Edition Release 11.1.0.6.0 - Production

--------------------------------------------------------------
#!/bin/bash
export
n s e
i ce
b l el
fer a
set echo on
a n s
o n -tr
a n users temporary
drop user jfv cascade;
create user jfv identified by jfv default

h s
a deฺ
tablespace
tablespace temp;
e )
s Gui
grant connect, resource, dbaฺm
r ฺ
toi jfv; ent
d g p t u d
connect jfv/jfv
a@ this S
b i
i a (ja purge;
drop table orders
u se
A b to (order_id_char varchar2(50) primary key,
ni o
create table orders
ntoorder_total number, customer_name varchar2(300));
e A
J o s begin
for i in 1..500000 loop
insert into orders
values(i,100,'aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
aaaaaaaaaaaaaaaaaaaaaaa');
end loop;
commit;
end;
/
begin
insert into orders
values(500000+i,100,'aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaa');
end loop;
commit;

end;
/
drop table employees purge;

drop table job_history purge;
create table employees (employee_id number primary key, name

varchar2(500));
begin
insert into employees
values(i,'aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
aaaaaaaaaaaaaaaaaaa');
n s e
end loop;
i ce
commit;
b l el
end;
/
fer a
a n s
o n -tr
create table job_history (employee_id number, job varchar2(500));
begin
s an
) h a deฺ
insert into job_history
r ฺ e s Gui
ฺ m i ent
values(mod(i,1000),'aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
d g p t u d
a@ this
aaaaaaaaaaaaaaaaaaaaaaaaaaaaa');S
end loop; b i
commit;
i a (ja use
end;
A b to
o
ni index job_history_empid_indx on job_history(employee_id);
/
ntocreate
e A
J o s
drop table old purge;
drop table new purge;
create table old(name varchar2(10), other varchar2(500));

create table new(name varchar2(10), other varchar2(500));
begin
insert into old
values(i,'aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
aaaaaaaaaaaaaaaaaaa');
end loop;
commit;
end;
/
begin

insert into new
values(500000+i,'aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
aaaaaaaaaaaaaaaaaaaaaaaaaa');
end loop;
commit;
end;
/
EOF
9) Connect as the JFV user from a SQL*Plus session and stay connected in that session
n s e
until further notice. In the session, set SQL*Plus timings on and flush your
ice
environment again before starting the second case. You can use the set timing
b l el
on command and the flush.sql script for this.
f er a
an s
[oracle@edrsr33p1-orcl Common_Mistakes]$ sqlplus jfv/jfv
- t r
n 26 21:06:33
n oMar
SQL*Plus: Release 11.1.0.6.0 - Production on Wed
s a
a eฺ
2008
) h i d
Copyright (c) 1982, 2007, Oracle. es
i r t Gu reserved.
ฺ Allnrights
g p ฺm ude
Connected to: d S t
i a @ h i s
Oracle Database 11g
b t
Enterprise Edition Release 11.1.0.6.0 -
Production
a (ja useOracle Label Security, OLAP, Data Mining
With the iPartitioning,
A
and Real to Testing options
b Application
o
ni set timing on
ntoSQL>
e A SQL> @flush
J o s
System altered.
Elapsed: 00:00:00.33
System altered.
Elapsed: 00:00:07.12
SQL>
--------------------------------------------------------------


10) The second case you analyze is a join case. From your SQL*Plus session, execute the
following query and note the time it takes to complete:
SELECT count(*)
FROM job_history jh, employees e
WHERE
substr(to_char(e.employee_id),1)=substr(to_char(jh.employee_id),1);
SQL> @join
SQL>
SQL> SELECT count(*)
2 FROM job_history jh, employees e
3 WHERE substr(to_char(e.employee_id),1) =
substr(to_char(jh.employee_id),1);
COUNT(*)
----------
n s e
499500
i ce
b l el
Elapsed: 00:00:03.03
SQL> fe r a
a n s
n
--------------------------------------------------------------
o -tr
set timing on a n
set echo on
h s
a deฺ
e )
s Gui
r ฺ
mi e dent
SELECT count(*)
FROM
p
job_history jh, employees ฺ
g Stu
d
WHERE substr(to_char(e.employee_id),1) =
b i a@ this
substr(to_char(jh.employee_id),1);
i a (ja use
b to fixtthe
11) Before trying
A o previous statement, flush your environment again using the
i o
flush.sql
n script from your SQL*Plus session.
t o
e An SQL> @flush
Jos
System altered.
Elapsed: 00:00:00.12
SQL> alter system flush buffer_cache;
System altered.
Elapsed: 00:00:00.72
SQL>
--------------------------------------------------------------

12) How would you rewrite the previous query for better performance? Test your
solution. You should discuss this with your instructor.

SQL> @better_join
SQL>
SQL> set timing on
SQL> set echo on
SQL>
SQL> SELECT count(*)
2 FROM job_history jh, employees e

3 WHERE e.employee_id = jh.employee_id;
COUNT(*)
----------
499500
Elapsed: 00:00:00.70
SQL>
n s e
--------------------------------------------------------------
ice
b l el
set timing on
fe r a
set echo on
a n s
SELECT count(*)
o n -tr
an
FROM job_history jh, employees e
WHERE e.employee_id = jh.employee_id;
h s
a deฺ
e )
s Guagain i
r
13) Before analyzing the third case, flush yourฺ
i ent environment using the flush.sql
ฺ m
gp Stud
script from your SQL*Plus session.
d
SQL> @flush
i @ shared_pool;
aflush h is
b t
(ja use
SQL> alter system
b i a to
A
System altered.
o
t o ni
Elapsed: 00:00:00.11
A n SQL> alter system flush buffer_cache;
e
Jos System altered.
Elapsed: 00:00:00.23
SQL>
--------------------------------------------------------------

14) The third case you analyze is a simple predicate case. Still connected as the JFV user
from your SQL*Plus session, execute the following query and note the time it takes
to complete:
SELECT * FROM orders WHERE order_id_char = 1205;
SQL> @simple_predicate
SQL>
SQL> set timing on
SQL> set echo on

SQL>
SQL> SELECT * FROM orders WHERE order_id_char = 1205;
ORDER_ID_CHAR ORDER_TOTAL
-------------------------------------------------- -----------
CUSTOMER_NAME
--------------------------------------------------------------------
------------
1205 100
aaaaaaaaaaaa
aaaaaaaaaaaa
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
n s e
Elapsed: 00:00:07.62
i ce
SQL>
b l el
--------------------------------------------------------------
fe r a
a n s
set timing on
o n -tr
an
set echo on
SELECT * FROM orders WHERE order_id_chara=s1205; ฺ

s ) h
u i de
m i rฺein step
n Gflush your environment again
t 14,
ฺ
15) Before trying to fix the SELECT statement
gp Stud e
using the flush.sql script. d
b i a@ this
(ja flush
SQL> @flush
a
SQL> alter system
i u se shared_pool;
A b to
ni o
System altered.
t o
e An Elapsed: 00:00:00.13
Jos
System altered.
Elapsed: 00:00:00.20
SQL>
16) How would you rewrite the previous statement for better performance? Test your
SQL> @better_predicate
SQL>
SQL> set timing on
SQL> set echo on
SQL>
SQL> SELECT * FROM orders WHERE order_id_char = '1205';
ORDER_ID_CHAR ORDER_TOTAL
-------------------------------------------------- -----------
CUSTOMER_NAME

--------------------------------------------------------------------
------------
1205 100
aaaaaaaaaaaa
aaaaaaaaaaaa
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
Elapsed: 00:00:00.04
SQL>
--------------------------------------------------------------
set timing on
n s e
set echo on
i ce
b l el
SELECT * FROM orders WHERE order_id_char = '1205';
fe ra
a n s
n - tr
17) Before proceeding with the next analysis, flush your environment again using the
flush.sql script. o
s an
SQL> alter system flush shared_pool; ) h
SQL> @flush a deฺ
r ฺ e s Gui
System altered.
p ฺ mi dent
Elapsed: 00:00:00.11@d
g Stu
b
SQL> alter system i t h is
aflush buffer_cache;
i a (ja use
b
System altered.
A to
ni o
e A SQL>
J o s --------------------------------------------------------------

18) The fourth case is a UNION case. Execute the following query and note the time it
takes to complete:
select count(*)
from (select name from old union select name from new);
SQL> @union
SQL>
SQL> set timing on
SQL> set echo on
SQL>
SQL> select count(*)
2 from (select name from old union select name from new);
COUNT(*)

----------
1000000
Elapsed: 00:00:06.39
SQL>
19) Before investigating a better solution, flush your environment again using the
flush.sql script.
SQL> @flush
System altered.
Elapsed: 00:00:00.10
n s e
i ce
System altered.
b l el
Elapsed: 00:00:00.09
fe r a
SQL>
a n s
n -tr
20) How would you rewrite the previous statement for better performance? Test your
o
an
s
) h a deฺ
SQL> @better_union
SQL>
r ฺ e s Gui
SQL> set timing on
p ฺ mi dent
SQL> set echo on
d g Stu
a@ this
SQL>
b i
i a (ja name
2 from (select
u sefrom old union all select name from new);
A b
COUNT(*) to
i o
ton 1000000
----------
A n
o s e
J Elapsed: 00:00:00.42
SQL>
21) Execute the multiple_setup.sql script to set up the environment for this case.
SQL> @multiple_setup
SQL> create table myemp as select * from hr.employees;
Table created.
SQL> insert into myemp select * from myemp;

/
/
/
/
/
/
/
/
/
/

/
/
/
/
/
commit;
107 rows created.
SQL>
214 rows created.
SQL>
428 rows created.
SQL>
n s e
856 rows created.
i ce
b l el
SQL>
1712 rows created.
fer a
a n s
SQL>
n -tr
o
an
3424 rows created.
SQL>
h s
a deฺ
6848 rows created.
e )
s Gui
r ฺ
i ent
SQL>
p ฺ m
13696 rows created.
d g Stud
SQL> b i a@ this
i a (ja use
27392 rows created.
SQL> A
b to
o
ni rows created.
nto54784
e A
J o s SQL>
109568 rows created.
SQL>
SQL>
SQL>
SQL>
SQL>
SQL>
Commit complete.

SQL> insert into myemp select * from myemp;

commit;

SQL>
Commit complete.
SQL>
22) Execute the multiple1.sql script and note the total time it takes to execute.
SQL> @multiple1
SQL> set timing on
n s e
i ce
el
SQL>
SQL>
a b l
SQL> SELECT COUNT (*)
fe r
FROM myemp
a n s
WHERE salary < 2000;
n - tr
2 3
o
s an
) h a deฺ
COUNT(*)
r ฺ e s Gui
----------
p ฺ mi dent
0
d g Stu
b
Elapsed: 00:00:15.49 i a@ this
( j
SQL> SELECT COUNT a (*)se
FROM myemp
WHERE A salary to u 2000 AND 4000;
bia BETWEEN
o n i2o 3
nt COUNT(*)
e A ----------
J o s 5636096
Elapsed: 00:00:17.14
SQL> SELECT COUNT (*)
FROM myemp
WHERE salary>4000;
2 3
COUNT(*)
----------
8388608
Elapsed: 00:00:18.12
SQL>
23) How would you rewrite the statements found in multiple1.sql script for better
performance?
SQL> @multiple2
SQL> SELECT COUNT (CASE WHEN salary < 2000
2 THEN 1 ELSE null END) count1,

3 COUNT (CASE WHEN salary BETWEEN 2001 AND 4000
4 THEN 1 ELSE null END) count2,
5 COUNT (CASE WHEN salary > 4000
6 THEN 1 ELSE null END) count3
7 FROM myemp;
COUNT1 COUNT2 COUNT3

---------- ---------- ----------
0 5636096 8388608
Elapsed: 00:00:18.19
SQL>

SQL> exit
n s e
ce
eli
a b l
fe r
a n s
n -tr
25) You now analyze a fifth case that deals with database connections. Execute the
o
an
bad_connect.sh script from your terminal window connected as the oracle
s
user. Note the time it takes to complete.
) h a deฺ
r ฺ e
[oracle@edrsr33p1-orcl Common_Mistakes]$ G ui
s ./bad_connect.sh
ฺ
Wed Mar 26 21:56:42 GMT-7 2008
p mi dent
Wed Mar 26 21:57:07 GMT-7
[oracle@edrsr33p1-orcl d
g Stu
2008
a@ this
Common_Mistakes]$
b i
i a (ja use
--------------------------------------------------------------
A b
#!/bin/bash to
o
ni /home/oracle/solutions/Common_Mistakes
ntocd
e A
J o s export ORACLE_SID=orcl
export
STREAM_NUM=0
MAX_STREAM=500
date
while [ $STREAM_NUM -lt $MAX_STREAM ]; do
# one more
let STREAM_NUM="STREAM_NUM+1"
# start one more stream

sqlplus -s jfv/jfv @select.sql >> /tmp/bad_connect.log 2>&1
done
date
--------------------------------------------------------------
select count(*) from dba_users;

exit;
26) Analyze the bad_connect.sh script and try to find a better solution to enhance
the performance of that application. Test your solution. You should discuss this with
your instructor. n s e
i ce
b l el
[oracle@edrsr33p1-orcl Common_Mistakes]$ ./better_connect.sh
fe r a
Wed Mar 26 22:00:48 GMT-7 2008
a n s
-tr
Wed Mar 26 22:00:50 GMT-7 2008
o n
s an
--------------------------------------------------------------
) h a deฺ
#!/bin/bash
r ฺ e s Gui
p ฺ mi dent
d g Stu
i a@ this
b
i a (ja use
A b to
ni o
export
ntoPATH=/u01/app/oracle/product/11.1.0/db_1/bin:/bin:/usr/bin:/usr/loca
e A l/bin:/usr/X11R6/bin:/usr/java/jdk1.5.0_11/bin:/bin
J o s
date
sqlplus -s jfv/jfv @select2.sql >> /tmp/better_connect.log 2>&1
date
--------------------------------------------------------------
declare
c number;
begin
select count(*) into c from dba_users;
end loop;
end;
/
exit;

27) Clean up your environment by executing the cleanup_rest.sh script from your
terminal session.
[oracle@edrsr33p1-orcl Common_Mistakes]$ ./cleanup_rest.sh

2008
Connected to:
Production
n s e
i ce
SQL> SQL> SQL> SQL>
b l el
User dropped.
f er a
SQL> SQL> Disconnected from Oracle Database 11g Enterprise s
an Edition
- t r
on Data Mining
n
With the Partitioning, Oracle Label Security, OLAP,
a
[oracle@edrsr33p1-orcl Common_Mistakes]$as
) h i d eฺ
i r ฺes nt Gu
--------------------------------------------------------------
g p ฺm ude
#!/bin/bash d S t
@ s
a b ia
e thi
j
b ia ( to us
o AORACLE_HOME=/u01/app/oracle/product/11.1.0/db_1
n i
ntoexport
e A
J o s export
set echo on
EOF

n s e
i ce
b l el
fer a
a n s
n -tr
o
s an
) h a deฺ
r ฺ e s Gui
p ฺ mi dent
d g Stu
b i a@ this
i a (ja use
Ab to
o n io
nt
e A
J o s

Practice 3-1: Understanding Optimizer Decisions
In this practice, you try to understand optimizer decisions relating to which execution
plan to use. All the scripts needed for this practice can be found in your
$HOME/solutions/Trace_Event directory.
1) Execute the te_setup.sh script. This script executes the following query and
generates a trace file that contains all optimizer decisions.

SELECT ch.channel_class, c.cust_city, t.calendar_quarter_desc,
SUM(s.amount_sold) sales_amount
FROM sh.sales s,sh.times t,sh.customers c,sh.channels ch
WHERE s.time_id = t.time_id AND
s.cust_id = c.cust_id AND
s.channel_id = ch.channel_id AND
c.cust_state_province = 'CA' AND
ch.channel_desc IN ('Internet','Catalog') AND
n s e
i ce
el
t.calendar_quarter_desc IN ('1999-Q1','1999-Q2')
GROUP BY ch.channel_class, c.cust_city, t.calendar_quarter_desc;
a b l
[oracle@edrsr33p1-orcl Trace_Event]$ ./te_setup.sh
fe r
a n s
o n -tr
SQL*Plus: Release 11.1.0.6.0 - Production on Wed Apr 9 22:16:53 2008
Copyright (c) 1982, 2007, Oracle. All rights n

a reserved.
h s
a deฺ
e )
s Gui
r ฺ
miEdition t
Connected to:
p
Oracle Database 11g Enterpriseฺ d e nRelease 11.1.0.6.0 -
Production g
dOracle S t u
@ s
thi options
ia Testing
With the Partitioning, Label Security, OLAP, Data Mining
a b
and Real Application
j e
SQL> b ia ( to us
i o A system flush shared_pool;
SQL> alter
n
nto
e A System altered.
J o s SQL>
SQL> alter user sh identified by sh account unlock;
User altered.
SQL>
SQL> connect sh/sh
Connected.
SQL>
SQL> ALTER SESSION SET TRACEFILE_IDENTIFIER = 'MYOPTIMIZER';
Session altered.
SQL>
SQL> alter session set events '10053 trace name context forever,
level 1';
Session altered.
SQL>

Practice 3-1: Understanding Optimizer Decisions (continued)
SQL> SELECT ch.channel_class, c.cust_city, t.calendar_quarter_desc,
2 FROM sh.sales s,sh.times t,sh.customers c,sh.channels ch
3 WHERE s.time_id = t.time_id AND
4 s.cust_id = c.cust_id AND
5 s.channel_id = ch.channel_id AND
6 c.cust_state_province = 'CA' AND

7 ch.channel_desc IN ('Internet','Catalog') AND
8 t.calendar_quarter_desc IN ('1999-Q1','1999-Q2')
9 GROUP BY ch.channel_class, c.cust_city,
t.calendar_quarter_desc;
no rows selected
SQL>
SQL> exit
n s e
ice
b l el
fe ra
orcl_ora_1801_MYOPTIMIZER.trc
a n s
orcl_ora_1801_MYOPTIMIZER.trm
o n -tr
aandn interpret the important
[oracle@edrsr33p1-orcl Trace_Event]$
2) With the help of your instructor, see the generatedatraces eฺ purposes. Do not use it
hdemonstration
)
parts of the trace file. Note: This lab is onlysfor u i d
i
on your production system unless explicitly r ฺe askedt G
by Oracle Support Services.
a) The 10053 trace output is g p ฺmdownudinto
broken
ena number of sections that broadly
t through in evaluating a plan. These
doptimizerSgoes
@
reflect the stages that the i s
ia query,thparameters
( a b
stages are asjfollows: e used by the optimizer, base statistical
ia to
information, base u s
table access cost, join order and method computations,
b
A for special features, such as query transformations.
n i orecosting
to[oracle@edrsr33p1-orcl Trace_Event]$ cat myoptimizer.trc
e An
Jos Trace file
/u01/app/oracle/diag/rdbms/orcl/orcl/trace/orcl_ora_32643_MYOPTIMIZE
R.trc
Production
ORACLE_HOME = /u01/app/oracle/product/11.1.0/db_1
System name: Linux
Node name: edrsr33p1.us.oracle.com
Release: 2.6.9-55.0.0.0.2.ELsmp
Version: #1 SMP Wed May 2 14:59:56 PDT 2007
Machine: i686
Instance name: orcl
Redo thread mounted by this instance: 1
Oracle process number: 18
Unix process pid: 32643, image: [email protected] (TNS
V1-V3)

*** 2008-04-09 21:46:51.950
*** SESSION ID:(134.35423) 2008-04-09 21:46:51.950
*** CLIENT ID:() 2008-04-09 21:46:51.950
*** SERVICE NAME:(SYS$USERS) 2008-04-09 21:46:51.950
*** MODULE NAME:(SQL*Plus) 2008-04-09 21:46:51.950
*** ACTION NAME:() 2008-04-09 21:46:51.950
Registered qb: SEL$1 0xb7e0905c (PARSER)

---------------------
QUERY BLOCK SIGNATURE
---------------------
signature (): qb_name=SEL$1 nbfros=4 flg=0
fro(0): flg=4 objn=72254 hint_alias="C"@"SEL$1"
fro(1): flg=4 objn=72252 hint_alias="CH"@"SEL$1"
fro(2): flg=4 objn=72192 hint_alias="S"@"SEL$1"
fro(3): flg=4 objn=72250 hint_alias="T"@"SEL$1"
n s e
i ce
SPM: statement not found in SMB
b l el
DOP: Automatic degree of parallelism is disabled: Parameter.
PM: Considering predicate move-around in query block SEL$1 (#0)
fer a
**************************
a n s
Predicate Move-Around (PM)
o n -tr
an
**************************
OPTIMIZER INFORMATION
h s
a deฺ
******************************************
e )
s session i
u(sql_id=70fqjd9u1zk7c)
----- Current SQL Statement for rthisฺ
i ent G
-----
p ฺ m
SELECT ch.channel_class, gc.cust_city, d t.calendar_quarter_desc,
d S t u
FROM sh.sales b a@ thit,sh.customers
i
s,sh.times
s c,sh.channels ch
j a e
ia ( to= uc.cust_id
s
A b
s.cust_id AND
o n io c.cust_state_province = 'CA' AND

nt
e A ch.channel_desc IN ('Internet','Catalog') AND
J o s t.calendar_quarter_desc IN ('1999-Q1','1999-Q2')
GROUP BY ch.channel_class, c.cust_city, t.calendar_quarter_desc
*******************************************
Legend
The following abbreviations are used by optimizer trace.
CBQT - cost-based query transformation
JPPD - join predicate push-down
OJPPD - old-style (non-cost-based) JPPD
FPD - filter push-down
PM - predicate move-around
CVM - complex view merging
SPJ - select-project-join
SJC - set join conversion
SU - subquery unnesting
OBYE - order by elimination
ST - star transformation
CNT - count(col) to count(*) transformation
JE - Join Elimination
qb - query block
LB - leaf blocks
DK - distinct keys

LB/K - average number of leaf blocks per key
DB/K - average number of data blocks per key
CLUF - clustering factor
NDV - number of distinct values
Resp - response cost
Card - cardinality
Resc - resource cost

NL - nested loops (join)
SM - sort merge (join)
HA - hash (join)
CPUSPEED - CPU Speed
IOTFRSPEED - I/O transfer speed
IOSEEKTIM - I/O seek time
SREADTIM - average single block read time
MREADTIM - average multiblock read time
MBRC - average multiblock read count
n s e
MAXTHR - maximum I/O system throughput
i ce
SLAVETHR - average slave I/O throughput
b l el
dmeth - distribution method
1: no partitioning required
fer a
2: value partitioned
a n s
4: right is random (round-robin)
n -tr
o
an
128: left is random (round-robin)
8: broadcast right and partition left
16: broadcast left and partition right
h s
a deฺ
)
s Gui
32: partition left using partitioning of right
e
r ฺ
mi dent
64: partition right using partitioning of left
256: run the join in serial
p ฺ
g Stu
d
0: invalid distribution method
sel - selectivity
ptn - partition b i a@ this
i a (ja use
***************************************
b to
PARAMETERS USED BY THE OPTIMIZER
A
o n io
********************************
*************************************
nt
e A PARAMETERS WITH ALTERED VALUES
J o s ******************************
Compilation Environment Dump
_smm_min_size = 204 KB
_smm_max_size = 40960 KB
_smm_px_max_size = 102400 KB
Bug Fix Control Environment
*************************************
PARAMETERS WITH DEFAULT VALUES
******************************
optimizer_mode_hinted = false
optimizer_features_hinted = 0.0.0
parallel_execution_enabled = true
parallel_query_forced_dop = 0
parallel_dml_forced_dop = 0
parallel_ddl_forced_degree = 0
parallel_ddl_forced_instances = 0
_query_rewrite_fudge = 90
optimizer_features_enable = 11.1.0.6

_optimizer_search_limit = 5
optimizer_use_invisible_indexes = false
flashback_data_archive_internal_cursor = 0
_optimizer_extended_stats_usage_control = 240
fix 3834770 = 1
fix 3746511 = enabled
fix 2194204 = disabled
…
n s e
i ce
b l el
***************************************
fer a
PARAMETERS IN OPT_PARAM HINT
a n s
****************************
o n -tr
an
***************************************
Column Usage Monitoring is ON: tracking level = 1
***************************************
h s
a deฺ
e )
s Gui
r ฺ
i on equery
nt block SEL$1 (#0)
ฺ
Considering Query Transformations
p m d
**************************
d g Stu
Query transformations
b i a@ this
**************************
(QT)
CBQT: Validity
i a (jachecks
u sepassed for 70fqjd9u1zk7c.
SEL$1 A
b
CSE: Considering
tocommon sub-expression elimination in query block
o n io (#0)
*************************
nt Common Subexpression elimination (CSE)
e A
J o s *************************
CSE: CSE not performed on query block SEL$1 (#0).
OBYE: Considering Order-by Elimination from view SEL$1 (#0)
***************************
Order-by elimination (OBYE)
***************************
OBYE: OBYE bypassed: no order by to eliminate.
JE: Considering Join Elimination on query block SEL$1 (#0)
*************************
Join Elimination (JE)
*************************
SQL:******* UNPARSED QUERY IS *******
SELECT "CH"."CHANNEL_CLASS" "CHANNEL_CLASS","C"."CUST_CITY"
"CUST_CITY","T"."CALENDAR_QUARTER_DESC"
"CALENDAR_QUARTER_DESC",SUM("S"."AMOUNT_SOLD") "SALES_AMOUNT" FROM
"SH"."SALES" "S","SH"."TIMES" "T","SH"."CUSTOMERS"
"C","SH"."CHANNELS" "CH" WHERE "S"."TIME_ID"="T"."TIME_ID" AND
"S"."CUST_ID"="C"."CUST_ID" AND "S"."CHANNEL_ID"="CH"."CHANNEL_ID"
AND "C"."CUST_STATE_PROVINCE"='CA' AND
("CH"."CHANNEL_DESC"='Internet' OR "CH"."CHANNEL_DESC"='Catalog')
AND ("T"."CALENDAR_QUARTER_DESC"='1999-Q1' OR

"T"."CALENDAR_QUARTER_DESC"='1999-Q2') GROUP BY
"CH"."CHANNEL_CLASS","C"."CUST_CITY","T"."CALENDAR_QUARTER_DESC"
Query block SEL$1 (#0) unchanged
CNT: Considering count(col) to count(*) on query block SEL$1 (#0)
*************************
Count(col) to Count(*) (CNT)
*************************
CNT: COUNT() to COUNT(*) not done.
*************************
*************************
n s e
i ce
b l el
fer a
n s
a
n -tr
o
an
h s
a deฺ
query block SEL$1 (#0) unchanged
e )
s Gui
r ฺ
mi dent
Considering Query Transformations on query block SEL$1 (#0)
**************************
p ฺ
g Stu
d
Query transformations (QT)
b i a@ this
**************************
CSE: Considering common sub-expression elimination in query block
SEL$1 (#0)
i a (ja use
b to
*************************
A
o n io
Common Subexpression elimination (CSE)
*************************
nt
e A CSE: CSE not performed on query block SEL$1 (#0).
J o s query block SEL$1 (#0) unchanged

apadrv-start sqlid=8087006336042125548
:
call(in-use=62016, alloc=81864), compile(in-use=64064,
alloc=66868), execution(in-use=3376, alloc=4060)
*******************************************
Peeked values of the binds in SQL statement
*******************************************
CBQT: Considering cost-based transformation on query block SEL$1

(#0)
********************************
COST-BASED QUERY TRANSFORMATIONS
********************************
FPD: Considering simple filter push (pre rewrite) in query block
SEL$1 (#0)
FPD: Current where clause predicates "S"."TIME_ID"="T"."TIME_ID"
AND "S"."CUST_ID"="C"."CUST_ID" AND
"S"."CHANNEL_ID"="CH"."CHANNEL_ID" AND
"C"."CUST_STATE_PROVINCE"='CA' AND ("CH"."CHANNEL_DESC"='Internet'

OR "CH"."CHANNEL_DESC"='Catalog') AND
("T"."CALENDAR_QUARTER_DESC"='1999-Q1' OR "T"."C
OBYE: Considering Order-by Elimination from view SEL$1 (#0)

***************************
Order-by elimination (OBYE)
***************************
OBYE: OBYE bypassed: no order by to eliminate.
Considering Query Transformations on query block SEL$1 (#0)
**************************
Query transformations (QT)
**************************
CSE: Considering common sub-expression elimination in query block
SEL$1 (#0)
*************************
Common Subexpression elimination (CSE)
n s e
*************************
i ce
CSE: CSE not performed on query block SEL$1 (#0).
b l el
kkqctdrvTD-start on query block SEL$1 (#0)
kkqctdrvTD-start: :
fer a
n
a s
o n -tr
kkqctdrvTD-cleanup: transform(in-use=0, alloc=0)
s an:
h a deฺ
alloc=109196), execution(in-use=3376,) alloc=4060)
r ฺ e s Gui
kkqctdrvTD-end: ฺ m i ent
d g p
call(in-use=62016, alloc=81864),
t u d compile(in-use=106808,
a@ this S
b i
i a (ja set-join
SJC: Considering
u se conversion in query block SEL$1 (#1)
to (SJC)
b Conversion
*************************
A
i o
Set-Join
n
nto*************************
e A SJC: not performed
J o s CNT: Considering count(col) to count(*) on query block SEL$1 (#1)

*************************
Count(col) to Count(*) (CNT)
*************************
CNT: COUNT() to COUNT(*) not done.
*************************
*************************

PM: Considering predicate move-around in query block SEL$1 (#1)
**************************
Predicate Move-Around (PM)
**************************
PM: PM bypassed: Outer query contains no views.
PM: PM bypassed: Outer query contains no views.

kkqctdrvTD-start: :
kkqctdrvTD-cleanup: transform(in-use=0, alloc=0) :

n s e
kkqctdrvTD-end:
i ce
b l el
fer a
a n s
kkqctdrvTD-start: :
o n -tr
an
h s
a deฺ
Registered qb: SEL$1 0xb7e01c6c (COPY) SEL$1)
r ฺ e s Gui
---------------------
QUERY BLOCK SIGNATURE ฺ m i ent
---------------------
d g p t u d
signature(): NULL @
a is S
b i t h
***********************************
(ja ByuPlacement
Cost-Based Group
i a se
b to
***********************************
A
n i o
GBP: Checking validity of GBP for query block SEL$1 (#1)
nto(#1)
GBP: Checking validity of group-by placement for query block SEL$1
e A
J o s GBP: Bypassed: QB has disjunction.
kkqctdrvTD-end:
GBP: Applying transformation directives

GBP: Checking validity of group-by placement for query block SEL$1
(#1)
GBP: Bypassed: QB has disjunction.
JPPD: Considering Cost-based predicate pushdown from query block
SEL$1 (#1)
************************************
Cost-based predicate pushdown (JPPD)
************************************
kkqctdrvTD-start: :


kkqctdrvTD-end:
JPPD: Applying transformation directives

query block SEL$1 (#1) unchanged
FPD: Considering simple filter push in query block SEL$1 (#1)
"S"."TIME_ID"="T"."TIME_ID" AND "S"."CUST_ID"="C"."CUST_ID" AND
"S"."CHANNEL_ID"="CH"."CHANNEL_ID" AND
n s e
i ce
b l el
try to generate transitive predicate from check constraints for
fe r a
query block SEL$1 (#1)
a n s
o n -tr
finally: "S"."TIME_ID"="T"."TIME_ID" AND "S"."CUST_ID"="C"."CUST_ID"
an
AND "S"."CHANNEL_ID"="CH"."CHANNEL_ID" AND
h s
a deฺ
)
s Gui
e
r ฺ
kkoqbc: optimizing query block
p ฺ miSEL$1de(#1)
nt
d g Stu
:
b i
call(in-use=81716, a@ alloc=98240),
t h is compile(in-use=124640,
alloc=132448),( j a s e
execution(in-use=3376, alloc=4060)
b i a to u
i o A
kkoqbc-subheap (create addr=0xb7d47960)
n
toQUERY BLOCK TEXT
****************
A n
o s e ****************
J SELECT ch.channel_class, c.cust_city, t.calendar_quarter_desc,
s.channel_id = ch.channel_id
---------------------
QUERY BLOCK SIGNATURE
---------------------
signature (optimizer): qb_name=SEL$1 nbfros=4 flg=0
fro(0): flg=0 objn=72254 hint_alias="C"@"SEL$1"
fro(1): flg=0 objn=72252 hint_alias="CH"@"SEL$1"
fro(2): flg=0 objn=72192 hint_alias="S"@"SEL$1"
fro(3): flg=0 objn=72250 hint_alias="T"@"SEL$1"
-----------------------------
SYSTEM STATISTICS INFORMATION
-----------------------------
Using NOWORKLOAD Stats
CPUSPEED: 1263 millions instructions/sec

IOTFRSPEED: 4096 bytes per millisecond (default is 4096)
IOSEEKTIM: 10 milliseconds (default is 10)
***************************************
BASE STATISTICAL INFORMATION
***********************
Table Stats::
Table: CHANNELS Alias: CH
#Rows: 5 #Blks: 4 AvgRowLen: 40.00
Index Stats::
Index: CHANNELS_PK Col#: 1
LVLS: 0 #LB: 1 #DK: 5 LB/K: 1.00 DB/K: 1.00 CLUF: 1.00
***********************
Table Stats::
Table: CUSTOMERS Alias: C
n s e
Index Stats::
i ce
Index: CUSTOMERS_GENDER_BIX Col#: 4
b l el
Index: CUSTOMERS_MARITAL_BIX Col#: 6
fer a
a n s
Index: CUSTOMERS_PK Col#: 1
n -tr
o
an
LVLS: 1 #LB: 115 #DK: 55500 LB/K: 1.00 DB/K: 1.00 CLUF:
54405.00
Index: CUSTOMERS_YOB_BIX Col#: 5
h s
a deฺ
)
s Gui
e
r ฺ
mi dent
***********************
Table Stats::
p ฺ
g Stu
Table: TIMES Alias: T
d
Index Stats:: b i a@ this
i a (ja use
Index: TIMES_PK Col#: 1
b to
A
o n io
***********************
Table Stats::
nt
e A Table: SALES Alias: S (Using composite stats)
J o s #Rows: 918843 #Blks: 1769 AvgRowLen: 29.00

Index Stats::
Index: SALES_CHANNEL_BIX Col#: 4
USING COMPOSITE STATS
Index: SALES_CUST_BIX Col#: 2
35808.00
Index: SALES_PROD_BIX Col#: 1
1074.00
Index: SALES_PROMO_BIX Col#: 5
Index: SALES_TIME_BIX Col#: 3
1460.00
Access path analysis for SALES

***************************************
SINGLE TABLE ACCESS PATH
Single Table Cardinality Estimation for SALES[S]
Table: SALES Alias: S
Card: Original: 918843.000000 Rounded: 918843 Computed:
918843.00 Non Adjusted: 918843.00
Access Path: TableScan

Cost: 498.20 Resp: 498.20 Degree: 0
Cost_io: 481.00 Cost_cpu: 260685437
Resp_io: 481.00 Resp_cpu: 260685437
****** trying bitmap/domain indexes ******
Access Path: index (FullScan)
Index: SALES_CHANNEL_BIX
resc_io: 75.00 resc_cpu: 552508
ix_sel: 1.000000 ix_sel_with_filters: 1.000000
n s e
i ce
Index: SALES_CUST_BIX
b l el
resc_io: 503.00 resc_cpu: 10743684
fer a
a n s
n -tr
o
an
Index: SALES_PROD_BIX
h s
a deฺ
e )
s Gui
r ฺ
mi dent
Index: SALES_PROMO_BIX
p ฺ
g Stu
d
b i a@ this
i a (ja use
b to
Index: SALES_TIME_BIX
A
o n io
nt
e A Cost: 60.05 Resp: 60.05 Degree: 0
J o s Access Path: index (FullScan)

Index: SALES_PROMO_BIX
Cost: 58.03
Access path: Bitmap index - accepted

Cost: 2895.306181 Cost_io: 2869.400000 Cost_cpu:
392576474.936000 Sel: 1.000000
Not Believed to be index-only
****** finished trying bitmap/domain indexes ******
Best:: AccessPath: TableScan
Cost: 498.20 Degree: 1 Resp: 498.20 Card: 918843.00
Bytes: 0
Access path analysis for TIMES

***************************************
Single Table Cardinality Estimation for TIMES[T]
Table: TIMES Alias: T

Card: Original: 1826.000000 Rounded: 183 Computed: 182.60 Non
Adjusted: 182.60
Cost_io: 18.00 Cost_cpu: 2314640
Resp_io: 18.00 Resp_cpu: 2314640

Index: TIMES_PK
Cost: 18.15 Degree: 1 Resp: 18.15 Card: 182.60 Bytes: 0
n s e
Access path analysis for CUSTOMERS
i ce
***************************************
b l el
Single Table Cardinality Estimation for CUSTOMERS[C]
fer a
Table: CUSTOMERS Alias: C
a n s
-tr
Card: Original: 55500.000000 Rounded: 383 Computed: 382.76
n
o
an
Non Adjusted: 382.76
h s
a deฺ
Cost_io: 404.00 Cost_cpu: 32782460
e )
s Gui
r ฺ
mi dent
Resp_io: 404.00 Resp_cpu: 32782460
p ฺ
g Stu
d
b i a@ this
Index: CUSTOMERS_GENDER_BIX
i a (ja use
b to
A
o n io
Index: CUSTOMERS_MARITAL_BIX
nt
e A resc_io: 6.00 resc_cpu: 46329
J o s ix_sel: 1.000000 ix_sel_with_filters: 1.000000

Index: CUSTOMERS_PK
resc_io: 116.00 resc_cpu: 11926087
Index: CUSTOMERS_YOB_BIX
Cost: 2367.447569 Cost_io: 2364.560000 Cost_cpu: 43757572.166400
Sel: 1.000000

Cost: 406.16 Degree: 1 Resp: 406.16 Card: 382.76 Bytes:
0
Access path analysis for CHANNELS

***************************************
Single Table Cardinality Estimation for CHANNELS[CH]
Table: CHANNELS Alias: CH
Card: Original: 5.000000 Rounded: 2 Computed: 2.00 Non
Adjusted: 2.00
n s e
i ce
b l el
Index: CHANNELS_PK
fe r a
a n s
n - tr
o
an
h s
a deฺ
e )
s Gu2 i
r ฺ
i ent 286
Grouping column cardinality [CHANNEL_CL]
Grouping column cardinality ฺ[mCUST_CITY]
d g
Grouping column cardinalityp t u d
[CALENDAR_Q] 2
a@ this S
***************************************
b i
i a (ja usANDe
A b
OPTIMIZER
to
STATISTICS COMPUTATIONS
o n io PLANS
***************************************
GENERAL
nt ***************************************
e A
J o s Considering cardinality-based initial join order.
Permutations for Starting Table :0
Join order[1]: CHANNELS[CH]#0 TIMES[T]#1 CUSTOMERS[C]#2
SALES[S]#3
***************
Now joining: TIMES[T]#1
***************
NL Join
Outer table: Card: 2.00 Cost: 3.00 Resp: 3.00 Degree: 1 Bytes:
21
Inner table: TIMES Alias: T
NL Join: Cost: 37.31 Resp: 37.31 Degree: 1
Index: TIMES_PK

Best NL cost: 37.31

resc: 37.31 resc_io: 37.00 resc_cpu: 4659106
resp: 37.31 resp_io: 37.00 resc_cpu: 4659106

Join Card: 365.200000 = = outer (2.000000) * inner (182.600000) *
sel (1.000000)
Join Card - Rounded: 365 Computed: 365.20
Grouping column cardinality [CHANNEL_CL] 2
Grouping column cardinality [ CUST_CITY] 286
Grouping column cardinality [CALENDAR_Q] 2
Best:: JoinMethod: NestedLoop
n s e
***************
i ce
Now joining: CUSTOMERS[C]#2
b l el
***************
NL Join
fer a
n
Outer table: Card: 365.20 Cost: 37.31 Resp: 37.31 Degree: 1
a s
Bytes: 37
n -tr
o
an
Access path analysis for CUSTOMERS
Inner table: CUSTOMERS Alias: C
h s
a deฺ
)
s Gui
e
r ฺ
mi dent
Cost_io: 146936.00 Cost_cpu: 11970256947
p ฺ
Resp_io: 146936.00 Resp_cpu: 11970256947
g Stu
d
b i a@ this
i a (ja use
b to
A
o n io
nt
e A Index: CUSTOMERS_MARITAL_BIX
J o s resc_io: 6.00 resc_cpu: 46329

Index: CUSTOMERS_PK
resc_io: 116.00 resc_cpu: 11926087
Index: CUSTOMERS_YOB_BIX
NL Join : Cost: 1498.02 Resp: 1498.02 Degree: 1
Cost_io: 1497.00 Cost_cpu: 15421408
Resp_io: 1497.00 Resp_cpu: 15421408

15980832052.096001 Sel: 1.000000
Best NL cost: 147725.92


Join Card: 139783.448276 = = outer (365.200000) * inner
(382.758621) * sel (1.000000)
Best:: JoinMethod: NestedLoop
n s e
Bytes: 63
i ce
b l el
***************
Now joining: SALES[S]#3
fer a
***************
a n s
NL Join
n -tr
o
an
Outer table: Card: 139783.45 Cost: 147725.92 Resp: 147725.92
Degree: 1 Bytes: 63
h s
a deฺ
Inner table: SALES Alias: S
e )
s Gui
r ฺ
mi dent
p ฺ
g Stu
d
Cost_io: 2538743.82 Cost_cpu: 1313377131608
b i a@ this
Resp_io: 2538743.82 Resp_cpu: 1313377131608
i a (ja use
Access Path: index (AllEqJoinGuess)
b to
Index: SALES_CHANNEL_BIX
A
o n io
nt
e A NL Join : Cost: 1686111.78 Resp: 1686111.78 Degree: 1
J o s Cost_io: 1684549.00 Cost_cpu: 23682093020

Resp_io: 1684549.00 Resp_cpu: 23682093020
Index: SALES_CUST_BIX
Cost_io: 286719.00 Cost_cpu: 13252268345
Resp_io: 286719.00 Resp_cpu: 13252268345
Index: SALES_TIME_BIX
Cost_io: 286719.00 Cost_cpu: 13112485345
Resp_io: 286719.00 Resp_cpu: 13112485345
38233905490.990532 Sel: 0.000000

(918843.000000) * sel (0.000000)

Outer table: CUSTOMERS Alias: C
resc: 147725.92 card 139783.45 bytes: 63 deg: 1 resp:
147725.92
resc: 498.20 card: 918843.00 bytes: 21 deg: 1 resp: 498.20
using dmeth: 2 #groups: 1
n s e
SORT ressource Sort statistics
i ce
Sort width: 238 Area size: 208896 Max Area size:
b l el
41943040
Degree: 1
fer a
Blocks to Sort: 1370 Row size: 80 Total Rows:
a n s
139783
n -tr
o
an
Initial runs: 2 Merge passes: 1 IO Cost / pass: 744
Total IO sort cost: 2114 Total CPU sort cost: 156539686
Total Temp space used: 21423000
h s
a deฺ
SORT ressource )
s Gui
Sort statistics
e
r ฺ
mi dent
41943040
p ฺ
g Stu
Degree:
d 1
918843 b i a@ this
i a (ja use
b to
Total IO sort cost: 5899
A
Total CPU sort cost: 929421655
o n io Total Temp space used: 66626000

SM join: Resc: 156308.78 Resp: 156308.78 [multiMatchCost=0.00]
nt
e A SM Join
J o s SM cost: 156308.78
resp: 156308.78 resp_io: 155430.00 resp_cpu: 13316903726
resc: 147725.92 card 139783.45 bytes: 63 deg: 1 resp:
147725.92
Cost per ptn: 1944.36 #ptns: 1
hash_area: 124 (max=10240) buildfrag: 1280 probefrag: 3702
ppasses: 1
Hash join: Resc: 150168.48 Resp: 150168.48 [multiMatchCost=0.00]
HA Join
HA cost: 150168.48
GROUP BY sort
GROUP BY adjustment factor: 0.500000
GROUP BY cardinality: 572.000000, TABLE cardinality: 3116.000000

41943040
Degree: 1
3116

Best:: JoinMethod: Hash
Bytes: 84
***********************
Best so far: Table#: 0 cost: 3.0020 card: 2.0000 bytes: 42
Table#: 1 cost: 37.3075 card: 365.2000 bytes: 13505
Table#: 2 cost: 147725.9191 card: 139783.4483
bytes: 8806329
n s e
Table#: 3 cost: 150169.5886 card: 3115.5952 bytes:
i ce
261744
b l el
***********************
Join order[2]: CHANNELS[CH]#0 TIMES[T]#1 SALES[S]#3
fe r a
CUSTOMERS[C]#2
a n s
n - tr
o
an
***************
Now joining: SALES[S]#3
***************
h s
a deฺ
NL Join
e )
s Gui
r ฺ
mi dent
Bytes: 37
p ฺ
g Stu
d
b i a@ this
i a (ja use
b to
n
…
i oA
nto***********************
e A
J o s Join order[22]:
CHANNELS[CH]#0
SALES[S]#3 CUSTOMERS[C]#2 TIMES[T]#1
***************
Now joining: TIMES[T]#1
***************
NL Join
Bytes: 47
Cost_io: 797001.00 Cost_cpu: 115712978623
Resp_io: 797001.00 Resp_cpu: 115712978623
Access Path: index (UniqueScan)
Index: TIMES_PK
Cost_io: 50707.00 Cost_cpu: 894143044

Resp_io: 50707.00 Resp_cpu: 894143044
Access Path: index (AllEqUnique)
Index: TIMES_PK
Cost_io: 50707.00 Cost_cpu: 894143044

Resp_io: 50707.00 Resp_cpu: 894143044

(182.600000) * sel (0.000685)
n s e
i ce
b l el
Grouping column cardinality [ CUST_CITY]
Grouping column cardinality [CALENDAR_Q]
286
2
fer a
a n s
-tr
resc: 910.93 card 49822.22 bytes: 47 deg: 1 resp: 910.93
n
o
an
h s
a deฺ
SORT ressource )
s Gui
Sort statistics
e
r ฺ
mi dent
41943040
p ฺ
g Stu
Degree:
d 1
49822 b i a@ this
i a (ja use
b to
Total IO sort cost: 585
A
Total CPU sort cost: 59514576
o n io Total Temp space used: 6022000

nt
e A Sort width: 238 Area size: 208896 Max Area size:
J o s 41943040
Degree: 1
Blocks to Sort: 1 Row size: 28 Total Rows: 183
SM Join
SM cost: 1519.02
ppasses: 1
Outer table: TIMES Alias: T

Inner table: CUSTOMERS Alias: C
ppasses: 1
HA Join
HA cost: 929.92 swapped
Best:: JoinMethod: Hash
Cost: 929.92 Degree: 1 Resp: 929.92 Card: 6231.19 Bytes:
63
n s e
***************
i ce
Now joining: CHANNELS[CH]#0
b l el
***************
NL Join
f a
er 1
Outer table: Card: 6231.19 Cost: 929.92 Resp: 929.92
an s
Degree:
Bytes: 63
n -tr
o
an
Access path analysis for CHANNELS
Inner table: CHANNELS Alias: CH
h s
a deฺ
)
s Gui
e
r ฺ
mi dent
Cost_io: 7655.00 Cost_cpu: 593732024
p ฺ
Resp_io: 7655.00 Resp_cpu: 593732024
g Stu
d
Access Path: index (UniqueScan)
Index: CHANNELS_PK
b i a@ this
i a (ja use
b to
A
o n io Cost_io: 7134.00 Cost_cpu: 460548636
Resp_io: 7134.00 Resp_cpu: 460548636
nt
e A Access Path: index (AllEqUnique)
J o s Index: CHANNELS_PK
Cost_io: 7134.00 Cost_cpu: 460548636
Resp_io: 7134.00 Resp_cpu: 460548636

Join Card: 3115.595241 = = outer (6231.190483) * inner (2.000000) *
sel (0.250000)

41943040
Degree: 1

41943040
Degree: 1
n s e
i ce
b l el
SM Join
fer a
SM cost: 1033.25
a n s
-
n tr
o
an
h s
a deฺ
e )
s Gui
r ฺ
mi dent
p ฺ
g Stu
d
1 b i a@ this
hash_area: 124 (max=10240) buildfrag: 58 probefrag: 1 ppasses:
i a (ja use
b to
Outer table: CHANNELS Alias: CH
A
o n io
nt
e A resc: 929.92 card: 6231.19 bytes: 63 deg: 1 resp: 929.92
J o s using dmeth: 2 #groups: 1

hash_area: 124 (max=10240) buildfrag: 1 probefrag: 58 ppasses:
1
HA Join
HA cost: 933.46 swapped
GROUP BY sort
41943040
Degree: 1
3116

Join order aborted: cost > best plan cost
***********************
(newjo-stop-1) k:0, spcnt:0, perm:22, maxperm:2000
*********************************
Number of join permutations tried: 22
*********************************
(newjo-save) [1 3 2 0 ]
41943040
Degree: 1
3116
n s e
i ce
b l el
Trying or-Expansion on query block SEL$1 (#1)
fer a
Transfer Optimizer annotations for query block SEL$1 (#1)
a n s
id=0 frofand predicate="C"."CUST_STATE_PROVINCE"='CA'
n -tr
o
an
id=0 frofkksm[i] (sort-merge/hash)
predicate="S"."CUST_ID"="C"."CUST_ID"
h s
a deฺ
id=0 frosand (sort-merge/hash) predicate="S"."CUST_ID"="C"."CUST_ID"
e )
s Gui
r ฺ
mi dent
predicate="S"."TIME_ID"="T"."TIME_ID"
p ฺ
id=0 frosand (sort-merge/hash) predicate="S"."TIME_ID"="T"."TIME_ID"
g Stu
d
id=0 frofand predicate="T"."CALENDAR_QUARTER_DESC"='1999-Q1' OR
b i a@ this
"T"."CALENDAR_QUARTER_DESC"='1999-Q2'
i a (ja use
predicate="S"."CHANNEL_ID"="CH"."CHANNEL_ID"
b to
id=0 frosand (sort-merge/hash)
A
o n io
predicate="S"."CHANNEL_ID"="CH"."CHANNEL_ID"
id=0 frofand predicate="CH"."CHANNEL_DESC"='Catalog' OR
nt
e A "CH"."CHANNEL_DESC"='Internet'
J o s GROUP BY adjustment factor: 1.000000

Final cost for query block SEL$1 (#1) - All Rows Plan:
Best join order: 16
Cost: 934.5672 Degree: 1 Card: 3116.0000 Bytes: 261744
Resc: 934.5672 Resc_io: 906.0000 Resc_cpu: 432900898
Resp: 934.5672 Resp_io: 906.0000 Resc_cpu: 432900898
kkoqbc-subheap (delete addr=0xb7d47960, in-use=105564, alloc=107812)
kkoqbc-end:
:
kkoqbc: finish optimizing query block SEL$1 (#1)

apadrv-end
:
Starting SQL statement dump

user_id=92 user_name=SH module=SQL*Plus action=
sql_id=70fqjd9u1zk7c plan_hash_value=1647000731 problem_type=3
----- Current SQL Statement for this session (sql_id=70fqjd9u1zk7c)
-----

sql_text_length=473
sql=SELECT ch.channel_class, c.cust_city, t.calendar_quarter_desc,
n s e
i ce
b l el
fer a
a n s
sql=AND
o n -tr
an
h s
a deฺ
)
s Gui
e
r ฺ
mi dent
----- Explain Plan Dump -----
----- Plan Table -----
p ฺ
g Stu
d
============
Plan Table b i a@ this
============(ja se
b i a to u
----------------------------------------------------+---------------
o A
i | Operation
--------------------+---------------+
nto||n Id | Name | Rows | Bytes
e A Cost | Time | Pstart| Pstop |
J o s ----------------------------------------------------+---------------
--------------------+---------------+
| 0 | SELECT STATEMENT | | |
| 935 | | | |
| 1 | HASH GROUP BY | | 572 | 47K
| 935 | 00:00:12 | | |
| 2 | HASH JOIN | | 3116 | 256K
| 933 | 00:00:12 | | |
| 3 | TABLE ACCESS FULL | CHANNELS | 2 | 42
| 3 | 00:00:01 | | |
| 4 | HASH JOIN | | 6231 | 383K
| 930 | 00:00:12 | | |
| 5 | PART JOIN FILTER CREATE | :BF0000 | 183 | 2928
| 18 | 00:00:01 | | |
| 6 | TABLE ACCESS FULL | TIMES | 183 | 2928
| 18 | 00:00:01 | | |
| 7 | HASH JOIN | | 49K | 2287K
| 911 | 00:00:11 | | |
| 8 | TABLE ACCESS FULL | CUSTOMERS| 383 | 9958
| 406 | 00:00:05 | | |

| 9 | PARTITION RANGE JOIN-FILTER | | 897K | 18M
| 498 | 00:00:06 | :BF0000| :BF0000|
| 10 | TABLE ACCESS FULL | SALES | 897K | 18M
| 498 | 00:00:06 | :BF0000| :BF0000|
----------------------------------------------------+---------------
--------------------+---------------+
Predicate Information:
----------------------
2 - access("S"."CHANNEL_ID"="CH"."CHANNEL_ID")
3 - filter(("CH"."CHANNEL_DESC"='Catalog' OR
"CH"."CHANNEL_DESC"='Internet'))
4 - access("S"."TIME_ID"="T"."TIME_ID")
6 - filter(("T"."CALENDAR_QUARTER_DESC"='1999-Q1' OR
"T"."CALENDAR_QUARTER_DESC"='1999-Q2'))
7 - access("S"."CUST_ID"="C"."CUST_ID")
8 - filter("C"."CUST_STATE_PROVINCE"='CA')
n s e
i ce
Content of other_xml column
b l el
===========================
db_version : 11.1.0.6
fer a
parse_schema : SH
a n s
plan_hash : 1647000731
n -tr
o
an
plan_hash_2 : 94088042
Outline Data:
/*+
h s
a deฺ
BEGIN_OUTLINE_DATA
e )
s Gui
r ฺ
mi dent
IGNORE_OPTIM_EMBEDDED_HINTS
p ฺ
OPTIMIZER_FEATURES_ENABLE('11.1.0.6')
g Stu
d
DB_VERSION('11.1.0.6')
ALL_ROWS
b i a@ this
OUTLINE_LEAF(@"SEL$1")
i a (ja use
FULL(@"SEL$1" "C"@"SEL$1")
b to
FULL(@"SEL$1" "S"@"SEL$1")
A
o n io FULL(@"SEL$1" "T"@"SEL$1")
FULL(@"SEL$1" "CH"@"SEL$1")
nt
e A LEADING(@"SEL$1" "C"@"SEL$1" "S"@"SEL$1" "T"@"SEL$1"
J o s "CH"@"SEL$1")
USE_HASH(@"SEL$1" "S"@"SEL$1")
USE_HASH(@"SEL$1" "T"@"SEL$1")
USE_HASH(@"SEL$1" "CH"@"SEL$1")
SWAP_JOIN_INPUTS(@"SEL$1" "T"@"SEL$1")
SWAP_JOIN_INPUTS(@"SEL$1" "CH"@"SEL$1")
USE_HASH_AGGREGATION(@"SEL$1")
END_OUTLINE_DATA
*/
Optimizer state dump:

optimizer_mode_hinted = false
optimizer_features_hinted = 0.0.0
parallel_execution_enabled = true
parallel_query_forced_dop = 0
parallel_dml_forced_dop = 0
parallel_ddl_forced_degree = 0
parallel_ddl_forced_instances = 0
_query_rewrite_fudge = 90
optimizer_features_enable = 11.1.0.6

…

fix 3834770 = 1

fix 2194204 = disabled
n s e
Query Block Registry:
i ce
SEL$1 0xb7e0905c (PARSER) [FINAL]
b l el
:
fe r a
a n s
o n -tr
End of Optimizer State Dump n
a ======================
====================== END SQL StatementaDump s eฺ
[oracle@edrsr33p1-orcl Trace_Event]$ ) h i d
i r ฺes nt Gu
3) Execute the te_cleanup.sh g p ฺm clean
script totu deup your environment for this lab.
d S ./te_cleanup.sh
@ Trace_Event]$
i a i s
ab se th
[oracle@edrsr33p1-orcl
SQL*Plus:ia
( j u
A b t o
Release 11.1.0.6.0 - Production on Wed Apr 9 22:18:01 2008
o n io
t
e An
Jos Connected to:
Production
SQL>
SQL> exit;
[oracle@edrsr33p1-orcl Trace_Event]$

n s e
i ce
b l el
fer a
a n s
n -tr
o
s an
) h a deฺ
r ฺ e s Gui
p ฺ mi dent
d g Stu
b i a@ this
i a (ja use
Ab to
o n io
nt
e A
J o s

Practice 4-1: Using Different Access Paths
In this practice, you explore various access paths the optimizer can use, and compare
them. You have the possibility of exploring 16 different scenarios, each of which are self-
contained. All scripts needed for this lab can be found in your
$HOME/solutions/Access_Paths directory.
1) Case 1: Connected as the oracle user from a terminal session, execute the
ap_setup.sh script.
[oracle@edrsr33p1-orcl Access_Paths]$ ./ap_setup.sh

n s e
ice
Connected to:
b l el
fe ra
Production
a n s
n - tr
o
SQL>
s an
SQL> alter user sh identified by sh account
) h i d eฺ
a unlock;
User altered. i r ฺes nt Gu
g p ฺm ude
SQL> d S t
@ s
thi
SQL> grant dba to a
a b i sh;
j e
ia ( to us
Grant succeeded.
b
i o A
n
SQL>
toSQL> exit;
A n Disconnected from Oracle Database 11g Enterprise Edition Release
o s e 11.1.0.6.0 - Production
J With the Partitioning, Oracle Label Security, OLAP, Data Mining
[oracle@edrsr33p1-orcl Access_Paths]$
2) In the same terminal session, connect as the SH user in the SQL*Plus session.
[oracle@edrsr33p1-orcl Access_Paths]$ sqlplus sh/sh
Connected to:
Production
SQL>

Practice 4-1: Using Different Access Paths (continued)
3) Unless otherwise notified, stay connected to your SQL*Plus session as the SH user.
Execute the idx_setup.sql script to set up your environment for case 1.
SQL> @idx_setup
SQL>
SQL> drop table mysales purge;
drop table mysales purge

*
ERROR at line 1:
SQL>
SQL> create table mysales as select * from sh.sales;
Table created.
n s e
i ce
SQL>
b l el
SQL> insert into mysales select * from mysales;
fe r a
a n s
n - tr
o
an
SQL> commit;
Commit complete. h s
a deฺ
e )
s Gui
r ฺ
SQL>
SQL>
p ฺ mi dent
d
SQL> insert into mysales gselectS*tufrom mysales;
b i a@ this
(ja use
b i a to
A
SQL> commit;
o
ni complete.
toCommit
A n
o s e SQL>
J SQL> insert into mysales select * from mysales;
SQL> commit;
Commit complete.
SQL>
SQL> commit;
Commit complete.
SQL>

SQL> commit;
Commit complete.
SQL>
SQL> insert into mysales values (0,0,sysdate,0,0,0,0);
1 row created.
SQL> commit;
Commit complete.
SQL>
n s e
SQL> exec dbms_stats.gather_schema_stats('SH');
ice
b l el
fe ra
SQL>
a n s
o n
4) Set up your session with the following commands: set timing on, set
-tr
an
autotrace trace only, and set linesize 200. After this, execute the
s
h a deฺ
select * from mysales where prod_id=0; query. What do you observe?
)
r ฺ e s Gtable. ui
p ฺ mi dent
a) Basically, there are no indexes on the MYSALES
b) .The only possibility ford goptimizer
the
S tuis to use the full table scan to retrieve only
one row. You can a see@ s takes a long time.
that the iscan
b i t h
i a (jaon use
SQL> @with_and_without_index
SQL> set
SQL> A
b echo
to
o
ni set timing on
ntoSQL>
e A SQL> set autotrace traceonly
J o s SQL> set linesize 200 pagesize 1000

SQL>
System altered.
Elapsed: 00:00:00.61
System altered.
Elapsed: 00:00:15.04
SQL>
SQL> select * from mysales where prod_id=0;
Elapsed: 00:00:28.85
Execution Plan
----------------------------------------------------------
Plan hash value: 3597614299

--------------------------------------------------------------------
---------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)|
Time |
--------------------------------------------------------------------
---------
| 0 | SELECT STATEMENT | | 402K| 11M| 40249 (2)|

00:08:03 |
|* 1 | TABLE ACCESS FULL| MYSALES | 402K| 11M| 40249 (2)|
00:08:03 |
--------------------------------------------------------------------
---------
Predicate Information (identified by operation id):

---------------------------------------------------
n s e
1 - filter("PROD_ID"=0)
i ce
b l el
Statistics
fe r a
----------------------------------------------------------
a n s
421 recursive calls
n - tr
o
an
0 db block gets
141606 consistent gets
141529 physical reads
h s
a deฺ
0 redo size
e )
s Gui
r ฺ
mi dent
790 bytes sent via SQL*Net to client
p ฺ
420 bytes received via SQL*Net from client
g Stu
d
2 SQL*Net roundtrips to/from client
b i a@ this
5 sorts (memory)
0 sorts (disk)
i a (ja use
1 rows processed
Ab to
o n io
SQL>
SQL> set timing off
nt
e A SQL> set autotrace off
J o s SQL>
5) How do you enhance the performance of the query in step 4? Implement your
solution.
SQL> @create_mysales_index
SQL> set echo on
SQL>
SQL> create index mysales_prodid_idx on mysales(prod_id) nologging
compute statistics;
Index created.
SQL>
6) Execute the same query again and verify that performance is enhanced now.
a) After implementing the index, the optimizer can use it to speed up the query
execution time. You can see a dramatic improvement in performance.
SQL> @with_and_without_index
SQL> set echo on

SQL>
SQL> set timing on
SQL> set autotrace traceonly
SQL> set linesize 200 pagesize 1000
SQL>
System altered.
Elapsed: 00:00:00.29
System altered.
Elapsed: 00:00:01.71
SQL>
n s e
SQL> select * from mysales where prod_id=0;
i ce
b l el
Elapsed: 00:00:00.88
fer a
Execution Plan
a n s
o n -tr
----------------------------------------------------------
an
h s
a deฺ
--------------------------------------------------------------------
------------------------------
e )
s| NameGui
Bytes | Cost (%CPU)| Time ฺm|i
| Id | Operation r ฺ n t | Rows |
d g p tude
--------------------------------------------------------------------
i a @ his S |
------------------------------
|
11M| 6079 (j(1)|ab 00:01:13
0 | SELECT STATEMENT
e t | | 402K|
a ACCESS
1 | biTABLE u sBY INDEX ROWID| MYSALES
|
11M| A
o
t 00:01:13 | | 402K|
o n
|* io2 | INDEX RANGE SCAN
6079 (1)|
| MYSALES_PRODID_IDX | 402K|
nt | 822 (1)| 00:00:10 |
e A
J o s --------------------------------------------------------------------
------------------------------

---------------------------------------------------
2 - access("PROD_ID"=0)
Statistics
----------------------------------------------------------
496 recursive calls
0 db block gets
104 consistent gets
21 physical reads
0 redo size
6 sorts (memory)
0 sorts (disk)

1 rows processed
SQL>
SQL> set timing off
SQL> set autotrace off
SQL>
7) Clean up your environment for case 1 by executing the idx_cleanup.sql script.

SQL> @idx_cleanup
SQL> set echo on
SQL>
SQL> drop table mysales purge;
Table dropped.
n s e
ce
SQL>
SQL>
eli
b l
8) Case 2: Drop all indexes currently created on the CUSTOMERS table except its
f e ra
primary key index.
t r a ns
SQL> @drop_customers_indexes
o n -
SQL> set termout off
s an
SQL>
) h a deฺ
ฺ e s Gui
--------------------------------------------------------------
r
p ฺ mi dent
set termout off
store set sqlplus_settings d g replace
S tu
buffer.sqlia @ his
save
timingja b replace
e t off verify off autotrace off feedback off
set
( off
s
heading
bia to u
spool Adait.sql
o n io
t SELECT 'drop index '||i.index_name||';'
e An FROM user_indexes i
Jos WHERE
AND
i.table_name = 'CUSTOMERS'
NOT EXISTS
(SELECT 'x'
FROM user_constraints c
WHERE c.index_name = i.index_name
AND c.table_name = i.table_name
AND c.status = 'ENABLED');
spool off
@dait
get buffer.sql nolist

@sqlplus_settings
set termout on

9) Execute the following query:
SELECT /*+ FULL(c) */ c.*
FROM customers c
WHERE cust_gender = 'M'
AND cust_postal_code = 40804
AND cust_credit_limit = 10000;
What do you observe?

a) The optimizer uses a full table scan and the cost for this query is relatively high.
SQL> @query00
SQL> set echo on
SQL>
SQL> set timing on
SQL>
SQL>
set autotrace traceonly
set linesize 200 pagesize 1000
n s e
i ce
el
SQL>
SQL> SELECT /*+ FULL(c) */ c.*
a b l
2 FROM customers c
fe r
3 WHERE cust_gender = 'M'
a n s
-tr
4 AND cust_postal_code = 40804
5 AND cust_credit_limit = 10000
o n
6 /
s an
6 rows selected.
) h a deฺ
r ฺ e s Gui
Elapsed: 00:00:00.29
p ฺ mi dent
Execution Plan d g Stu
i a@ this
----------------------------------------------------------
b
i a (ja use
A b to
--------------------------------------------------------------------
i o
to|n Id | Operation
-----------
| Name | Rows | Bytes | Cost (%CPU)|
A n Time |
o s e --------------------------------------------------------------------
J -----------
| 0 | SELECT STATEMENT | | 6 | 1080 | 407 (1)|
00:00:05 |
|* 1 | TABLE ACCESS FULL| CUSTOMERS | 6 | 1080 | 407 (1)|
00:00:05 |
--------------------------------------------------------------------
-----------

---------------------------------------------------
1 - filter(TO_NUMBER("CUST_POSTAL_CODE")=40804 AND
"CUST_CREDIT_LIMIT"=10000 AND "CUST_GENDER"='M')
Statistics
----------------------------------------------------------
1067 recursive calls
0 db block gets

1461 physical reads
116 redo size
27 sorts (memory)
0 sorts (disk)
6 rows processed
SQL>
SQL> set timing off
SQL>
10) Create three B*-tree indexes on the following CUSTOMERS table columns:
cust_gender
n s e
cust_postal_code i ce
cust_credit_limit
b l el
fe ra
SQL> @create_cust_gender_index
a n s
SQL>
SQL>
set echo on
n - tr
o
an
SQL> CREATE INDEX cust_cust_gender_idx
2
3
ON customers(cust_gender)
NOLOGGING COMPUTE STATISTICS; h s
a deฺ
e )
s Gui
r ฺ
Index created.
p ฺ mi dent
SQL> d g Stu
SQL> set echo onbi
a@ this
SQL> @create_cust_postal_code_index
SQL>
i a (ja use
A b INDEXtocust_cust_postal_code_idx
SQL> CREATE
n o
i3 NOLOGGING
2 ON customers(cust_postal_code)
nt o COMPUTE STATISTICS;
e A
J o s Index created.
SQL>
SQL> @create_cust_credit_limit_index
SQL> set echo on
SQL>
SQL> CREATE INDEX cust_cust_credit_limit_idx
2 ON customers(cust_credit_limit)
3 NOLOGGING COMPUTE STATISTICS;
Index created.
SQL>
SQL> @list_customers_indexes
SQL> set echo on
SQL>
SQL> SELECT ui.table_name
2 , decode(ui.index_type
3 ,'NORMAL', ui.uniqueness
4 ,ui.index_type) AS index_type

5 , ui.index_name
6 FROM user_indexes ui
7 WHERE ui.table_name = 'CUSTOMERS'
8 ORDER BY ui.table_name
9 , ui.uniqueness desc;
TABLE_NAME INDEX_TYPE
INDEX_NAME
------------------------------ --------------------------- ---------
---------------------
CUSTOMERS UNIQUE
CUSTOMERS_PK
CUSTOMERS NONUNIQUE
CUST_CUST_GENDER_IDX
CUSTOMERS NONUNIQUE
CUST_CUST_CREDIT_LIMIT_IDX
n s e
CUSTOMERS NONUNIQUE
i ce
CUST_CUST_POSTAL_CODE_IDX
b l el
SQL>
fe ra
a n s
11) Start monitoring all the CUSTOMERS indexes.
o n -tr
SQL> @start_monitoring_indexes
s an
SQL> set echo on
) h a deฺ
ui
SQL>
r ฺ
SQL> ALTER INDEX CUSTOMERS_PK MONITORINGe s USAGE;
G
p ฺ mi dent
Index altered.
d g Stu
SQL>
b i a@ this
( j
SQL> ALTER INDEXa s e
CUST_CUST_POSTAL_CODE_IDX MONITORING USAGE;
bia to u
Index A
altered.
o n io
nt SQL>
e A SQL> ALTER INDEX CUST_CUST_GENDER_IDX MONITORING USAGE;
J o s
Index altered.
SQL>
SQL> ALTER INDEX CUST_CUST_CREDIT_LIMIT_IDX MONITORING USAGE;
Index altered.
SQL>
SQL> @show_index_usage
SQL> set echo on
SQL>
SQL> set linesize 200
SQL>
SQL> select * from v$object_usage;
INDEX_NAME TABLE_NAME MON

USE START_MONITORING END_MONITORING
------------------------------ ------------------------------ --- --
- ------------------- -------------------

CUSTOMERS_PK CUSTOMERS YES NO
04/09/2008 14:52:43
CUST_CUST_POSTAL_CODE_IDX CUSTOMERS YES NO
04/09/2008 14:52:43
CUST_CUST_GENDER_IDX CUSTOMERS YES NO
04/09/2008 14:52:43
CUST_CUST_CREDIT_LIMIT_IDX CUSTOMERS YES NO

04/09/2008 14:52:43
SQL>

SELECT /*+ INDEX(c) */ c.*
FROM customers c
n s e
i ce
b l el
fe r a
a) The optimizer chooses to use only one index to do a fast full scan. The cost is
a n s
lower than the full table scan.
n - tr
o
an
SQL> @query01
SQL> set echo on
SQL> h s
a deฺ
SQL> set timing on
e )
s Gui
r ฺ
p
SQL> set linesize 200 pagesize ฺ mi1000dent
SQL> d g Stu
i a@c t*/
SQL> SELECT /*+ INDEX(c)
b h isc.*
2 FROM
3 WHERE (
a ja use = 'M'
customers
i cust_gender
to
b cust_postal_code
o
i6 /
5 A
4 AND
AND cust_credit_limit
= 40804
= 10000
t o n
A n
o s e 6 rows selected.
J
Elapsed: 00:00:00.04
Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
-------------------------------------
| Id | Operation | Name |
Rows | Bytes | Cost (%CPU)| Time |
--------------------------------------------------------------------
-------------------------------------
| 0 | SELECT STATEMENT | |
6 | 1080 | 218 (1)| 00:00:03 |
|* 1 | TABLE ACCESS BY INDEX ROWID| CUSTOMERS |
6 | 1080 | 218 (1)| 00:00:03 |
|* 2 | INDEX FULL SCAN | CUST_CUST_POSTAL_CODE_IDX |
89 | | 134 (1)| 00:00:02 |

--------------------------------------------------------------------
-------------------------------------

---------------------------------------------------
1 - filter("CUST_CREDIT_LIMIT"=10000 AND "CUST_GENDER"='M')

2 - filter(TO_NUMBER("CUST_POSTAL_CODE")=40804)
Statistics
----------------------------------------------------------
395 recursive calls
3 db block gets
384 consistent gets
132 physical reads
n s e
604 redo size
i ce
b l el
fe r a
6 sorts (memory)
a n s
0 sorts (disk)
n - tr
o
an
6 rows processed
SQL>
h s
a deฺ
SQL> set timing off
e )
s Gui
r ฺ
mi dent
SQL>
p ฺ
g Stu
d
b i a@ this*/ c.*
SELECT /*+ INDEX_COMBINE(c)
FROM (ja c us=e'M'
customers
i a
b
WHERE cust_gender
to
AND Acust_postal_code
n i o = 40804
to AND cust_credit_limit = 10000;
e An What do you observe?

Jos a) This time the optimizer uses multiple indexes and combines them to access the
table. However, the cost is higher than that from the previous step as well as the
full table scan.
SQL> @query02
SQL> set echo on
SQL>
SQL> set timing on
SQL>
SQL> SELECT /*+ INDEX_COMBINE(c) */ c.*
2 FROM customers c
6 /
6 rows selected.

Elapsed: 00:00:00.06
Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
-------------------------------------------
| Id | Operation | Name
| Rows | Bytes | Cost (%CPU)| Time |
--------------------------------------------------------------------
-------------------------------------------
| 0 | SELECT STATEMENT |
| 6 | 1080 | 466 (1)| 00:00:06 |
|* 1 | TABLE ACCESS BY INDEX ROWID | CUSTOMERS
| 6 | 1080 | 466 (1)| 00:00:06 |
n s e
| 2 | BITMAP CONVERSION TO ROWIDS |
i ce
| | | | |
b l el
|
|
3 |
|
BITMAP AND
| | |
|
fer a
| 4 | BITMAP CONVERSION FROM ROWIDS|
a n s
| | | | |
n -tr
o
an
|* 5 | INDEX RANGE SCAN |
CUST_CUST_CREDIT_LIMIT_IDX | | | 14 (0)| 00:00:01 |
| 6 |
h
BITMAP CONVERSION FROM ROWIDS| s
a deฺ
| | | |
e )
s Gui
|
r ฺ
mi dent
|* 7 | INDEX RANGE SCAN | CUST_CUST_GENDER_IDX
| | | 51 ฺ
(0)| 00:00:01 |
p
g Stu
d
--------------------------------------------------------------------
b i a@ this
-------------------------------------------
i a (ja use(identified by operation id):

Predicate Information
b to
---------------------------------------------------
A
i o
nton 15 -- filter(TO_NUMBER("CUST_POSTAL_CODE")=40804)
e A access("CUST_CREDIT_LIMIT"=10000)
J o s 7 - access("CUST_GENDER"='M')
Statistics
----------------------------------------------------------
3 recursive calls
7 db block gets
894 consistent gets
81 physical reads
1020 redo size
0 sorts (memory)
0 sorts (disk)
6 rows processed
SQL>
SQL> set timing off
SQL>

14) Confirm the list of indexes that were accessed in this case.
SQL> @show_index_usage
SQL> set echo on
SQL>
SQL>
SQL> select * from v$object_usage;
INDEX_NAME TABLE_NAME MON

USE START_MONITORING END_MONITORING
------------------------------ ------------------------------ --- --
- ------------------- -------------------
CUSTOMERS_PK CUSTOMERS YES NO
04/09/2008 14:52:43
CUST_CUST_POSTAL_CODE_IDX CUSTOMERS YES
n s e
YES 04/09/2008 14:52:43
i ce
CUST_CUST_GENDER_IDX
YES 04/09/2008 14:52:43
CUSTOMERS YES
b l el
CUST_CUST_CREDIT_LIMIT_IDX CUSTOMERS
fe a
rYES
YES 04/09/2008 14:52:43
a n s
SQL> o n -tr
SQL>
s an
15) Case 3: Drop all the CUSTOMERS indexes except ) h aits primary
i d eฺkey index. After this,
r
make sure you create a concatenated index
s the following
ฺeon G u CUSTOMERS columns,
i en t
and in the order mentioned here:pฺm d
cust_gender d g S t u
@ s
cust_credit_limit
j a b ia
e thi
b i a ( to us
cust_postal_code
i A termout off
o set
to n
SQL>
e An SQL>
SQL> @create_gender_limit_code_index
Jos SQL>
SQL>
set echo on
SQL> CREATE INDEX cust_gender_limit_code_idx

2 ON customers(cust_gender,cust_credit_limit,cust_postal_code)
Index created.
SQL>

FROM customers c
a) The optimizer uses your concatenated index, and the resulting cost is by far the
best compared to the previous steps.

SQL> @query01
SQL> set echo on
SQL>
SQL> set timing on
SQL>
SQL> SELECT /*+ INDEX(c) */ c.*
2 FROM customers c
6 /
6 rows selected.
n s e
Elapsed: 00:00:00.01
i ce
b l el
Execution Plan
fer a
----------------------------------------------------------
a n s
n -tr
o
an
--------------------------------------------------------------------
--------------------------------------
h s
a deฺ |
e )
s Gui
r ฺ
mi dent
--------------------------------------------------------------------
p ฺ
--------------------------------------
g Stu
| 0 | SELECT STATEMENT
d | |
7 | 1260 | 18
b a@ this
(0)| 00:00:01 |
i
(ja use
| 1 | TABLE ACCESS BY INDEX ROWID| CUSTOMERS |
7 | 1260 |
|* 2 | b i a18 (0)| 00:00:01 |
to
INDEX RANGE SCAN | CUST_GENDER_LIMIT_CODE_IDX |
6 |
io A | 12 (0)| 00:00:01 |
t o n
--------------------------------------------------------------------
A n --------------------------------------
o s e
J Predicate Information (identified by operation id):
---------------------------------------------------
2 - access("CUST_GENDER"='M' AND "CUST_CREDIT_LIMIT"=10000)

filter(TO_NUMBER("CUST_POSTAL_CODE")=40804)
Statistics
----------------------------------------------------------
1 recursive calls
0 db block gets
22 consistent gets
14 physical reads
0 redo size
0 sorts (memory)
0 sorts (disk)
6 rows processed

SQL>
SQL> set timing off
SQL>


FROM customers c
a) The query is almost the same as in the previous step, but the predicate is removed.
The optimizer can still use the concatenated index, but the resulting cost is much e
higher because cust_credit_limit is not very selective. n s
ice
SQL> @query03
b l el
SQL> set echo on
fe ra
SQL>
a n s
n - tr
SQL> set timing on
o
s an
SQL>
) h a deฺ
SQL>
2
FROM customers c
r ฺ e s Gui
= 'M'i nt
3 WHERE cust_gender
p ฺ m d e
4
5
AND
/ d g Stu
cust_credit_limit = 10000
b i a@ this
i a (ja use
4101 rows selected.
A
Elapsed: to
b 00:00:00.08
ni o
ntoExecution Plan
e A ----------------------------------------------------------
J o s Plan hash value: 2871279522
--------------------------------------------------------------------
--------------------------------------
--------------------------------------------------------------------
--------------------------------------
3469 | 609K| 3454 (1)| 00:00:42 |
3469 | 609K| 3454 (1)| 00:00:42 |
|* 2 | INDEX RANGE SCAN | CUST_GENDER_LIMIT_CODE_IDX |
3469 | | 12 (0)| 00:00:01 |
--------------------------------------------------------------------
--------------------------------------

---------------------------------------------------

2 - access("CUST_GENDER"='M' AND "CUST_CREDIT_LIMIT"=10000)
Statistics
----------------------------------------------------------
1 recursive calls
0 db block gets
0 physical reads
0 redo size
0 sorts (memory)
0 sorts (disk)
4101 rows processed
n s e
i ce
SQL>
b l el
SQL>
fer a
SQL> set timing off
a n s
SQL>
o n -tr
s an
) h a deฺ
FROM customers c
r ฺ e s Gui
mi dent
AND cust_postal_code = 40804;
p ฺ
g Stu
What do you observe? d
a) You replacedja b ia@ this
( u s e is used and with
cust_credit_limit cust_postal_code, which has
b ia to
better selectivity. The index the resulting cost is better.
i A
o @query04
n
SQL>
toSQL> set echo on
e An SQL>
Jos
SQL> set timing on
SQL>
2 FROM customers c
5 /
75 rows selected.
Elapsed: 00:00:00.02
Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
--------------------------------------

--------------------------------------------------------------------
--------------------------------------
45 | 8100 | 133 (1)| 00:00:02 |

45 | 8100 | 133 (1)| 00:00:02 |
|* 2 | INDEX RANGE SCAN | CUST_GENDER_LIMIT_CODE_IDX |
45 | | 87 (0)| 00:00:02 |
--------------------------------------------------------------------
--------------------------------------

---------------------------------------------------
n s e
2 - access("CUST_GENDER"='M')
i ce
filter(TO_NUMBER("CUST_POSTAL_CODE")=40804)
b l el
fe ra
Statistics
a n s
-
----------------------------------------------------------
n tr
o
an
1 recursive calls
0 db block gets
196 consistent gets
h s
a deฺ
101 physical reads
e )
s Gui
r ฺ
mi dent
0 redo size
p ฺ
g Stu
d
b i a@ this
0 sorts (memory)
i a (ja use
0 sorts (disk)
A b to
75 rows processed
o n io
SQL>
nt
e A SQL> set timing off
J o s SQL> set autotrace off

SQL>

FROM customers c
WHERE cust_postal_code = 40804
a) The leading part of the concatenated index is no longer part of the query.
However, the optimizer is still able to use the index by doing a fast full index
scan. Nevertheless, the resulting cost is not the best.
SQL> @query05
SQL> set echo on
SQL>
SQL> set timing on
SQL>

2 FROM customers c
3 WHERE cust_postal_code = 40804
5 /
15 rows selected.
Elapsed: 00:00:00.02
Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
--------------------------------------
n s e
i ce
b l el
--------------------------------------------------------------------
--------------------------------------
fer a
a n s |
11 | 1980 | 185 (1)| 00:00:03 |
o n -tr
an
11 | 1980 | 185 (1)| 00:00:03 |
|* 2 | INDEX FULL SCAN
h s
a deฺ
| CUST_GENDER_LIMIT_CODE_IDX |
11 | | 173 (1)| 00:00:03 |
e )
s Gui
r ฺ
mi dent
--------------------------------------------------------------------
p ฺ
--------------------------------------
g Stuby operation id):
d
b i a@ this
(identified
---------------------------------------------------
i a (ja use
b to
2 - access("CUST_CREDIT_LIMIT"=10000)
A
n i o filter(TO_NUMBER("CUST_POSTAL_CODE")=40804 AND
nto"CUST_CREDIT_LIMIT"=10000)
e A
J o s Statistics
----------------------------------------------------------
1 recursive calls
0 db block gets
189 consistent gets
56 physical reads
0 redo size
0 sorts (memory)
0 sorts (disk)
15 rows processed
SQL>
SQL> set timing off
SQL>
SQL>

20) Case 4: Drop all the CUSTOMERS indexes except its primary key index. After this,
create three different bitmap indexes on the following columns of the CUSTOMERS
table:
cust_gender
cust_postal_code
cust_credit_limit
SQL>
SQL> @create_cust_gender_bindex
SQL> set echo on
SQL>
SQL> CREATE BITMAP INDEX cust_cust_gender_bidx ON
customers(cust_gender)
n s e
i ce
b l el
Index created.
fe r a
SQL>
a n s
SQL> @create_cust_postal_code_bindex
o n -tr
an
SQL> set echo on
SQL>
h s
a deฺ
SQL> CREATE BITMAP INDEX cust_cust_postal_code_bidx ON
customers(cust_postal_code)
e )
s Gui
r ฺ
mi dent
p ฺ
g Stu
Index created.
d
b i a@ this
(ja use
SQL>
SQL> setb a
SQL> @create_cust_credit_limit_bindex
i
echo onto
A
io CREATE BITMAP INDEX cust_cust_credit_limit_bidx ON
SQL>
t o n
SQL>
A n customers(cust_credit_limit)
o s e 2 NOLOGGING COMPUTE STATISTICS;

J Index created.
SQL>
SQL> @list_customers_indexes
SQL> set echo on
SQL>
SQL> SELECT ui.table_name
2 , decode(ui.index_type
3 ,'NORMAL', ui.uniqueness
4 ,ui.index_type) AS index_type
5 , ui.index_name
6 FROM user_indexes ui
7 WHERE ui.table_name = 'CUSTOMERS'
8 ORDER BY ui.table_name
9 , ui.uniqueness desc;
TABLE_NAME INDEX_TYPE
INDEX_NAME

------------------------------ --------------------------- ---------
---------------------
CUSTOMERS UNIQUE
CUSTOMERS_PK
CUSTOMERS BITMAP
CUST_CUST_GENDER_BIDX
CUSTOMERS BITMAP
CUST_CUST_CREDIT_LIMIT_BIDX
CUSTOMERS BITMAP
CUST_CUST_POSTAL_CODE_BIDX
SQL>

SELECT /*+ INDEX_COMBINE(c) */ c.*
FROM customers c
n s e
i ce
AND
AND
cust_postal_code = 40804
cust_credit_limit = 10000; b l el
fe r a
a n s
- tr
a) The optimizer uses only two bitmap indexes to solve this query. However, the
n
o
an
cost is not a good one. It is a little lesser than cost of the full table scan.
s
SQL> @query02
) h a deฺ
SQL> set echo on
r ฺ e s Gui
SQL>
p
SQL> set linesize 200 pagesize ฺ mi1000dent
SQL> set timing on d g Stu
i a@ this
b
SQL>
SQL> SELECT (
a ja INDEX_COMBINE(c)
u se
b i /*+
t o
*/ c.*
A cust_gender
2 FROM
3o WHERE
i
customers c
= 'M'
t o n 4 AND cust_postal_code = 40804
A n 5 AND cust_credit_limit = 10000
o s e 6 /
J
6 rows selected.
Elapsed: 00:00:00.01
Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
----------------------------------------
--------------------------------------------------------------------
----------------------------------------
6 | 1080 | 402 (1)| 00:00:05 |
|* 1 | TABLE ACCESS BY INDEX ROWID | CUSTOMERS |
6 | 1080 | 402 (1)| 00:00:05 |

| 2 | BITMAP CONVERSION TO ROWIDS| |
| | | |
| 3 | BITMAP AND | |
| | | |
|* 4 | BITMAP INDEX SINGLE VALUE| CUST_CUST_CREDIT_LIMIT_BIDX |
| | | |
|* 5 | BITMAP INDEX SINGLE VALUE| CUST_CUST_GENDER_BIDX |

| | | |
--------------------------------------------------------------------
----------------------------------------

---------------------------------------------------
n s e
i ce
b l el
Statistics
fe r a
----------------------------------------------------------
a n s
1 recursive calls
n - tr
o
an
0 db block gets
813 consistent gets
5 physical reads
h s
a deฺ
0 redo size
e )
s Gui
r ฺ
mi dent
p ฺ
g Stu
d
b i a@ this
0 sorts (memory)
0 sorts (disk)
i a (ja use
6 rows processed
A b to
o n io
SQL>
nt
e A SQL> set timing off
J o s SQL>
SQL>
create two bitmap indexes on the following columns of the CUSTOMERS table:
cust_year_of_birth
cust_credit_limit
SQL>
SQL> @create_cust_year_of_birth_bindex
SQL> set echo on
SQL>
SQL> CREATE BITMAP INDEX cust_cust_year_of_birth_bidx
2 ON customers(cust_year_of_birth)
Index created.
SQL>

SQL> set echo on
SQL>
SQL> CREATE BITMAP INDEX cust_cust_credit_limit_bidx ON
customers(cust_credit_limit)
Index created.
SQL>

SELECT /*+ INDEX_COMBINE(c) */ c.*
FROM customers c
WHERE c.cust_year_of_birth = 1953
OR c.cust_credit_limit = 10000;
n s e
What do you observe? li ce
b l e
a) The query uses an OR construct. The optimizer can use both bitmap indexes.
f e ra
However, this resulting cost is not the best.
t r a ns
SQL> @query06
o n -
SQL> set echo on
SQL>
s an
) h a deฺ
SQL> set timing on
r ฺ e s Gui
SQL> p ฺ mi dent
d g Stu*/ c.*
SQL> SELECT /*+ INDEX_COMBINE(c)
2 FROM
i
customers
b a@c this
4 OR a ( ja use
3 WHERE c.cust_year_of_birth = 1953
5 / bi
c.cust_credit_limit = 10000
A to
o
ni rows selected.
nt o7158
eA
Jos
Elapsed: 00:00:00.12
Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
-----------------------------------------
--------------------------------------------------------------------
-----------------------------------------
| 7585 | 1333K| 581 (1)| 00:00:07 |
| 1 | TABLE ACCESS BY INDEX ROWID | CUSTOMERS
| 7585 | 1333K| 581 (1)| 00:00:07 |
| 2 | BITMAP CONVERSION TO ROWIDS|
| | | | |
| 3 | BITMAP OR |
| | | | |

|* 4 | BITMAP INDEX SINGLE VALUE| CUST_CUST_CREDIT_LIMIT_BIDX
| | | | |
|* 5 | BITMAP INDEX SINGLE VALUE| CUST_CUST_YEAR_OF_BIRTH_BIDX
| | | | |
--------------------------------------------------------------------
-----------------------------------------

---------------------------------------------------
4 - access("C"."CUST_CREDIT_LIMIT"=10000)
5 - access("C"."CUST_YEAR_OF_BIRTH"=1953)
Statistics
----------------------------------------------------------
n s e
1 recursive calls
i ce
0 db block gets
b l el
3 physical reads
fe r a
0 redo size
a n s
o n -tr
an
0 sorts (memory)
h s
a deฺ
0 sorts (disk)
e )
s Gui
7158 rows processed r ฺ
i ent
p ฺ m d
SQL>
SQL> set timing off @ d g S t u
SQL> set autotraceb i aoff this
SQL>
i a (ja use
SQL>
A b to
o
ni 6: Drop all the CUSTOMERS indexes except its primary key index. After this,
24) Case
t o
e An create three different bitmap indexes on the following columns of the CUSTOMERS
Jos
table:
cust_year_of_birth
cust_credit_limit
cust_postal_code
SQL>
SQL> @create_cust_year_of_birth_bindex
SQL> set echo on
SQL>
SQL> CREATE BITMAP INDEX cust_cust_year_of_birth_bidx
2 ON customers(cust_year_of_birth)
Index created.
SQL>
SQL> set echo on

SQL>
Index created.
SQL>
SQL> @create_cust_postal_code_bindex
SQL> set echo on
SQL>
SQL> CREATE BITMAP INDEX cust_cust_postal_code_bidx ON
customers(cust_postal_code)
Index created.
n s e
ice
SQL>
b l el
fe r a
SELECT c.*
a n s
FROM customers c
WHERE (c.cust_year_of_birth = '1970' And c.cust_postal_code o n -tr =
40804 ) a n
AND NOT cust_credit_limit = 15000;
h a s eฺ
) i d
i r ฺes nt Gu
a) The query has a complex WHERE
g p ฺmclauseu e is well suited for using bitmap
dthat
d t
than the full tablei @cost.his S indexes and the resulting cost is better
indexes. The optimizer uses two bitmap
a
scan
( j a b et
i a u s
SQL> @query07
b
A echo ont
SQL> set o
i o
n set timing on
SQL>
ntoSQL>
e A SQL> set autotrace traceonly
J o s SQL>
SQL> SELECT c.*
2 FROM customers c
3 WHERE (c.cust_year_of_birth = '1970' And c.cust_postal_code =
40804 )
4 AND NOT cust_credit_limit = 15000
5 /
Elapsed: 00:00:00.03
Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
------------------------------------------
--------------------------------------------------------------------
------------------------------------------

| 1 | 180 | 133 (0)| 00:00:02 |
|* 1 | TABLE ACCESS BY INDEX ROWID | CUSTOMERS
| 1 | 180 | 133 (0)| 00:00:02 |
| 2 | BITMAP CONVERSION TO ROWIDS |
| | | | |
| 3 | BITMAP MINUS |
| | | | |
| 4 | BITMAP MINUS |
| | | | |
|* 5 | BITMAP INDEX SINGLE VALUE| CUST_CUST_YEAR_OF_BIRTH_BIDX
| | | | |
|* 6 | BITMAP INDEX SINGLE VALUE| CUST_CUST_CREDIT_LIMIT_BIDX
| | | | |
|* 7 | BITMAP INDEX SINGLE VALUE | CUST_CUST_CREDIT_LIMIT_BIDX
| | | | |
n s e
--------------------------------------------------------------------
i ce
------------------------------------------
b l el
fe r a
---------------------------------------------------
a n s
o n -tr
an
1 - filter(TO_NUMBER("C"."CUST_POSTAL_CODE")=40804)
5 - access("C"."CUST_YEAR_OF_BIRTH"=1970)
h s
a deฺ
7 - )
s Gui
access("CUST_CREDIT_LIMIT" IS NULL)
e
r ฺ
i ent
p ฺ m d
Statistics
d g S t u
b
1 recursivea@ calls
t h is
----------------------------------------------------------
i
a blocksegets
0 (jdb
b i
773a t o u gets
consistent
n i o A 30 physical reads
nto redo size
e A 2024 bytes sent via SQL*Net to client
J o s 420
2
bytes received via SQL*Net from client
SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
1 rows processed
SQL>
SQL> set timing off
SQL>
26) Make sure the optimizer can no longer use the bitmap index you created on the
cust_year_of_birth column.
a) The best solution is to render it invisible.
SQL> show parameter optimizer_use_invisible_indexes
NAME TYPE VALUE

------------------------------------ ----------- -------------------
-----------

optimizer_use_invisible_indexes boolean FALSE
SQL> alter index cust_cust_year_of_birth_bidx invisible;
Index altered.
SQL> select index_name, visibility from user_indexes where

table_owner='SH' and table_name='CUSTOMERS';
INDEX_NAME VISIBILIT
------------------------------ ---------
CUST_CUST_CREDIT_LIMIT_BIDX VISIBLE
CUST_CUST_YEAR_OF_BIRTH_BIDX INVISIBLE
CUST_CUST_POSTAL_CODE_BIDX VISIBLE
CUSTOMERS_PK VISIBLE
4 rows selected.
n s e
i ce
SQL>
b l el
fe r a
SELECT c.*
a n s
FROM customers c
WHERE (c.cust_year_of_birth = '1970' And c.cust_postal_code o n -tr =
40804 ) a n
AND NOT cust_credit_limit = 15000;
h a s eฺ
) i d
i r ฺes nt Gu
a) This is the same query as in the
g p ฺm
previous
u d e However, the optimizer can no
step.
d t
i a @ his S indexes.
longer find a good plan that uses bitmap
SQL> @query07 b
( j aon se t
ia to u
SQL> set echo
b
SQL>
i
SQL>o A timing on
set
n
toSQL> set autotrace traceonly
A n SQL>
o s e SQL> SELECT c.*
J 2 FROM customers c
3 WHERE (c.cust_year_of_birth = '1970' And c.cust_postal_code =
40804 )
4 AND NOT cust_credit_limit = 15000
5 /
Elapsed: 00:00:00.03
Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
-----------
Time |
--------------------------------------------------------------------
-----------
00:00:05 |

00:00:05 |
--------------------------------------------------------------------
-----------

---------------------------------------------------
1 - filter(TO_NUMBER("C"."CUST_POSTAL_CODE")=40804 AND
"C"."CUST_YEAR_OF_BIRTH"=1970 AND
"CUST_CREDIT_LIMIT"<>15000)
Statistics
----------------------------------------------------------
393 recursive calls
n s e
0 db block gets
i ce
b l el
0 physical reads
0 redo size
fer a
a n s
n -tr
o
an
6 sorts (memory)
0 sorts (disk)
h s
a deฺ
1 rows processed
e )
s Gui
r ฺ
i ent
SQL>
p ฺ m d
SQL> set timing off
d g S t u
SQL> b i a@ this
a
(jthe se query:
i a
28) Case 7: Execute
b u
following
to
o
SELECTA c.*
t o ni customers
FROM c
e An WHERE cust_id IN (88340,104590,44910);
Jos What do you observe?

a) The optimizer can use the CUSTOMERS primary key index to solve this query.
The cost is very low for the resulting plan.
SQL> @query08
SQL> set echo on
SQL>
SQL> set timing on
SQL>
SQL> SELECT c.*
2 FROM customers c
3 WHERE cust_id IN (88340,104590,44910)
4 /
Elapsed: 00:00:00.07
Execution Plan
----------------------------------------------------------

--------------------------------------------------------------------
-------------------------
| Id | Operation | Name | Rows | Bytes
| Cost (%CPU)| Time |
--------------------------------------------------------------------
-------------------------
| 0 | SELECT STATEMENT | | 3 | 540
| 7 (0)| 00:00:01 |
| 1 | INLIST ITERATOR | | |
| | |
| 2 | TABLE ACCESS BY INDEX ROWID| CUSTOMERS | 3 | 540
| 7 (0)| 00:00:01 |
|* 3 | INDEX UNIQUE SCAN | CUSTOMERS_PK | 3 |
| 4 (0)| 00:00:01 |
n s e
--------------------------------------------------------------------
i ce
-------------------------
b l el
fe r a
---------------------------------------------------
a n s
o n -tr
an
3 - access("CUST_ID"=44910 OR "CUST_ID"=88340 OR
"CUST_ID"=104590)
h s
a deฺ
e )
s Gui
Statistics r ฺ
i ent
ฺ m
----------------------------------------------------------
p d
d
2 recursive calls g S t u
3 db block
b i a@gets
9 consistent t h
getsis
a
4 (jphysical sereads
b i
560a redo u
to sent via SQL*Net to client
size
i o A2022 bytes
nton 420 bytes received via SQL*Net from client
e A 2 SQL*Net roundtrips to/from client
J o s 0
0
sorts (memory)
sorts (disk)
1 rows processed
SQL>
SQL> set timing off
SQL>
29) Case 8: Drop all the indexes on the CUSTOMERS table except its primary key index.
After this, create a concatenated B*-tree index on the following columns of the
CUSTOMERS table, and in the order here:
cust_last_name
cust_first_name
SQL>
SQL> @create_last_first_name_index
SQL> set echo on

SQL>
SQL> CREATE INDEX cust_last_first_name_idx
2 ON customers(cust_last_name,cust_first_name)
Index created.
SQL>

SELECT c.cust_last_name
, c.cust_first_name
FROM customers c;
a) The optimizer can use only the index without accessing the table itself. The n s e
resulting cost is very good. i ce
b l el
SQL> @query09
fe ra
SQL> set echo on
a n s
SQL>
n - tr
o
an
SQL> set timing on
SQL>
SQL>
set autotrace traceonly
h s
a deฺ
)
s Gui
c.cust_first_name irฺe
SQL>SELECT c.cust_last_name
2
3
,
FROM customers c
p ฺ m dent
4 / d g Stu
b i a@ this
i a (ja use
55500 rows selected.
A
Elapsed: to
b 00:00:00.19
ni o
ntoExecution Plan
e A ----------------------------------------------------------
J o s Plan hash value: 445338993
--------------------------------------------------------------------
-----------------------------
| Id | Operation | Name | Rows |
Bytes | Cost (%CPU)| Time |
--------------------------------------------------------------------
-----------------------------
| 0 | SELECT STATEMENT | | 55500 |
812K| 55 (2)| 00:00:01 |
| 1 | INDEX FAST FULL SCAN| CUST_LAST_FIRST_NAME_IDX | 55500 |
812K| 55 (2)| 00:00:01 |
--------------------------------------------------------------------
-----------------------------
Statistics
----------------------------------------------------------
1 recursive calls
0 db block gets

193 physical reads
0 redo size
0 sorts (memory)
0 sorts (disk)
55500 rows processed
SQL>
SQL>
SQL> set timing off
SQL>
SQL>
31) Case 9: Drop all the indexes on the CUSTOMERS table except its primary key index. n s e
i ce
After this, create two B*-tree indexes on the following columns of the CUSTOMERS
b l el
table:
fe ra
cust_last_name
a n s
cust_first_name
n - tr
o
s an
) h a deฺ
s Gui
SQL>
SQL> @create_last_name_index
r ฺ e
SQL> set echo on
p ฺ mi dent
SQL>
d g Stu
SQL> CREATE INDEX cust_cust_last_name_idx
i a@ STATISTICS;
t h is
2 ON customers(cust_last_name)
b
3 NOLOGGING
i a (ja use
COMPUTE
Index A b
created. to
ni o
ntoSQL>
e A SQL> @create_first_name_index
J o s SQL> set echo on

SQL>
SQL> CREATE INDEX cust_cust_first_name_idx
2 ON customers(cust_first_name)
Index created.
SQL>

SELECT /*+ INDEX_JOIN(c cust_cust_first_name_idx
cust_cust_last_name_idx) */ c.cust_last_name
, c.cust_first_name
FROM customers c;
a) Although the optimizer can use both the indexes, the resulting cost is not better
than the concatenated index case.

SQL> @query10
SQL> set echo on
SQL>
SQL> set timing on
SQL>
SQL> SELECT /*+ INDEX_JOIN(c cust_cust_first_name_idx
cust_cust_last_name_idx) */ c.cust_last_name
2 , c.cust_first_name
3 FROM customers c
4 /
Elapsed: 00:00:00.27
n s e
i ce
Execution Plan
b l el
----------------------------------------------------------
fer a
a n s
n -tr
--------------------------------------------------------------------
o
an
-------------------------------
| Id | Operation
| Name
h s
a deฺ
| Rows |
)
s Gui
--------------------------------------------------------------------
e
r ฺ
mi dent
-------------------------------
p ฺ
g Stu
| | 55500 |
812K| 511
d
(1)| 00:00:07 |
| 1 | VIEW
b i a@ this | index$_join$_001 | 55500 |
(ja use
812K| 511 (1)| 00:00:07 |
|* 2 |
| b i a
HASH JOIN
| to |
| | |
|
io
3 | A INDEX FAST FULL SCAN| CUST_CUST_FIRST_NAME_IDX | 55500 |
t o n
812K| 175 (1)| 00:00:03 |
A n | 4 | INDEX FAST FULL SCAN| CUST_CUST_LAST_NAME_IDX | 55500 |
o s e 812K| 178 (0)| 00:00:03 |

J --------------------------------------------------------------------
-------------------------------

---------------------------------------------------
2 - access(ROWID=ROWID)
Statistics
----------------------------------------------------------
1 recursive calls
0 db block gets
279 physical reads
0 redo size
0 sorts (memory)

0 sorts (disk)
SQL>
SQL>
SQL> set timing off

SQL>
SQL>
33) CASE 10: Drop all the indexes on the CUSTOMERS table except its primary key
index. Then, create one bitmap index on the following column of the CUSTOMERS
table:
cust_credit_limit
n s e
i ce
el
SQL>
a b l
fer
SQL> set echo on
a n s
-tr
SQL>
o n
2 NOLOGGING COMPUTE STATISTICS; s an
) h a deฺ
Index created.
r ฺ e s Gui
SQL> p ฺ mi dent
d g Stu
34) Execute the followinga
i @ his
query:
SELECT count(*)
j a b e t
credit_limit
FROM (
ia to u = 10000;
customers s
A b
WHERE cust_credit_limit
n i
What
o do you observe?
n to
e A a) Although cust_credit_limit is not a selective column, the COUNT
Jos operation on its bitmap index is very efficient.
SQL> @query11
SQL> set echo on
SQL>
SQL> set timing on
SQL>
SQL> SELECT count(*) credit_limit
2 FROM customers
3 WHERE cust_credit_limit = 10000
4 /
Elapsed: 00:00:00.00
Execution Plan
----------------------------------------------------------

--------------------------------------------------------------------
---------------------------------------
--------------------------------------------------------------------
---------------------------------------
1 | 4 | 1 (0)| 00:00:01 |
| 1 | SORT AGGREGATE | |
1 | 4 | | |
| 2 | BITMAP CONVERSION COUNT | |
6938 | 27752 | 1 (0)| 00:00:01 |
|* 3 | BITMAP INDEX SINGLE VALUE| CUST_CUST_CREDIT_LIMIT_BIDX |
| | | |
--------------------------------------------------------------------
---------------------------------------
n s e
i ce
b l el
---------------------------------------------------
fe r a
a n s
o n -tr
Statistics
s an
h a deฺ
----------------------------------------------------------
)
1 recursive calls
r ฺ e s Gui
3 consistent getsฺmi nt
0 db block gets
p
g Stu d e
d
2 physical reads
@ via
0 redo size
b
423 bytes i asent t h isSQL*Net to client
a received
420 (jbytes e via SQL*Net from client
sroundtrips
i a
b 0 sorts
2 u
to (memory) to/from client
SQL*Net
io A
t o n 0 sorts (disk)
A n 1 rows processed
o s e
J SQL>
SQL>
SQL> set timing off
SQL>
SQL>
create one B*-tree index on the following column of the CUSTOMERS table:
cust_credit_limit
SQL>
SQL> @create_cust_credit_limit_index
SQL> set echo on
SQL>
SQL> CREATE INDEX cust_cust_credit_limit_idx
2 ON customers(cust_credit_limit)

Index created.
SQL>

SELECT count(*) credit_limit

FROM customers
WHERE cust_credit_limit = 10000;
a) The optimizer can still use the index; however, this is less efficient compared to
the corresponding bitmap index from the previous case.
SQL> @query11
SQL> set echo on
n s e
i ce
el
SQL>
a b l
SQL> set timing on
fe r
a n s
-tr
SQL>
SQL> SELECT count(*) credit_limit
o n
an
2 FROM customers
3
4
WHERE cust_credit_limit = 10000
/ h s
a deฺ
e )
s Gui
r ฺ
Elapsed: 00:00:00.00
p ฺ mi dent
Execution Plan d g Stu
a@ this
Plan hash value:bi3421880850
----------------------------------------------------------
i a (ja use
A b to
--------------------------------------------------------------------
i o
----------------------------
nto|n Id | Operation | Name | Rows |
e A Bytes | Cost (%CPU)| Time |
J o s --------------------------------------------------------------------
----------------------------
4 | 14 (0)| 00:00:01 |
| 1 | SORT AGGREGATE | | 1 |
4 | | |
|* 2 | INDEX RANGE SCAN| CUST_CUST_CREDIT_LIMIT_IDX | 6938 |
27752 | 14 (0)| 00:00:01 |
--------------------------------------------------------------------
----------------------------

---------------------------------------------------
Statistics
----------------------------------------------------------
1 recursive calls
0 db block gets

14 consistent gets
13 physical reads
0 redo size
0 sorts (memory)
0 sorts (disk)
1 rows processed
SQL>
SQL>
SQL> set timing off
SQL>
SQL>
n s e
i ce
37) Case 12: Drop all the indexes on the CUSTOMERS table except its primary key index.
After this, create one B*-tree index on the following column of the CUSTOMERS b l el
fe r a
table: s
cust_last_name tra n
o n -
SQL> set termout off s an
SQL>
) h a deฺ
SQL> @create_last_name_index
r ฺ e s Gui
SQL> set echo on
SQL> p ฺ mi dent
d g Stu
SQL> CREATE INDEX cust_cust_last_name_idx
ia@ STATISTICS;
h is
2 ON customers(cust_last_name)
b
3 NOLOGGINGaCOMPUTE t
i a (j u se
A b
Index created. to
ni o
ntoSQL>
e A
J o s 38) Execute the following query:
SELECT cust_id, country_id
FROM customers
WHERE LOWER( cust_last_name) LIKE 'gentle';
a) Although there is an index, it cannot be used because its column is modified by a
function.
SQL> @query12
SQL> set echo on
SQL>
SQL> set timing on
SQL>
SQL> SELECT cust_id, country_id
2 FROM customers
3 WHERE LOWER( cust_last_name) LIKE 'gentle'
4 /

80 rows selected.
Elapsed: 00:00:00.01
Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
-----------
Time |
--------------------------------------------------------------------
-----------
n s e
00:00:05 |
i ce
b l el
00:00:05 |
--------------------------------------------------------------------
fe r a
-----------
a n s
o n -tr
an
---------------------------------------------------
h s
a deฺ
)
s Gui
1 - filter(LOWER("CUST_LAST_NAME")='gentle')
e
r ฺ
i ent
p ฺ m d
Statistics
d g S t u
b
1 recursivea@ calls
t h is
----------------------------------------------------------
i
a blocksegets
0 (jdb
b i
1464a t o u gets
consistent
nto redo size
J o s 475
7
0 sorts (memory)
0 sorts (disk)
80 rows processed
SQL>
SQL> set timing off
SQL>
39) How can you enhance the performance of the previous query without modifying the
statement itself? Implement your solution.
a) You can create a function-based index.
SQL> @create_lower_cust_last_name_index
SQL> set echo on
SQL>
SQL> CREATE INDEX lower_cust_last_name_idx ON
2 customers(LOWER(cust_last_name))
3 /

Index created.
SQL>
40) Check if your solution executes faster than in the case of the query in step 38.
SQL> @query12
SQL> set echo on
SQL>
SQL> set timing on
SQL>
SQL> SELECT cust_id, country_id
2 FROM customers
n s e
ce
3 WHERE LOWER( cust_last_name) LIKE 'gentle'
4 /
eli
a b l
80 rows selected.
fe r
a n s
Elapsed: 00:00:00.00
o n -tr
Execution Plan
s an
) h a deฺ
----------------------------------------------------------
s Gui
r ฺ e
p ฺ mi dent
--------------------------------------------------------------------
| Id | Operation d g Stu | Name

------------------------------------
|
@ s
Rows | Bytes | Cost
j a b ia
e thi
(%CPU)| Time |
i a ( u s
--------------------------------------------------------------------
------------------------------------
| 0A b t o
| SELECT STATEMENT | |
i o
to|n 1 | TABLE ACCESS BY INDEX ROWID| CUSTOMERS
555 | 17760 | 41 (0)| 00:00:01 |
A n |
o s e 555 | 17760 |
|* 2 |
41 (0)| 00:00:01 |
INDEX RANGE SCAN | LOWER_CUST_LAST_NAME_IDX |
J 222 | | 1 (0)| 00:00:01 |
--------------------------------------------------------------------
------------------------------------

---------------------------------------------------
2 - access(LOWER("CUST_LAST_NAME")='gentle')
Statistics
----------------------------------------------------------
24 recursive calls
0 db block gets
23 consistent gets
1 physical reads
0 redo size

0 sorts (memory)
0 sorts (disk)
80 rows processed
SQL>
SQL> set timing off

SQL>
SQL>
41) Case 13: Execute the iot_setup.sql script to set up the environment for this
case.
SQL> @iot_setup
SQL> set echo on
SQL>
n s e
SQL> CREATE table promotions_iot
i ce
2 (promo_id number primary key
b l el
3
4
, promo_name VARCHAR2(40)
, promo_subcategory VARCHAR2 (30)
fe ra
5 , promo_category VARCHAR2 (30)
a n s
6
7
, promo_cost NUMBER
, promo_begin_date DATE o n -tr
8 , promo_end_date DATE)
s an
9 ORGANIZATION INDEX
) h a deฺ
10 /
r ฺ e s Gui
Table created.
p ฺ mi dent
d g Stu
SQL>
a@ this
SQL> INSERT INTO ipromotions_iot
b
2 SELECT (
a ja usepromo_name, promo_subcategory, promo_category,
promo_id,
i
promo_cost,
to
b promotions
promo_begin_date, promo_end_date
A
i4o /
3 FROM
t o n
A n
o s e 503 rows created.
J SQL>
42) Execute the following two queries:

SELECT *
FROM promotions
WHERE promo_id > 300;
SELECT /*+ INDEX(promotions) */ *

FROM promotions
a) The first lets the optimizer decide the plan, and the best it can find is to do a full
table scan. Forcing the use of the index is not a good idea, as it takes longer to
execute.
SQL> @query13
SQL> set echo on
SQL>

SQL> set timing on
SQL>
SQL> SELECT *
2 FROM promotions
3 WHERE promo_id > 300

4 /
235 rows selected.
Elapsed: 00:00:00.02
Execution Plan
----------------------------------------------------------
n s e
i ce
--------------------------------------------------------------------
b l el
------------
| Id | Operation | Name | Rows | Bytes | Cost
fer a
(%CPU)| Time |
a n s
-tr
--------------------------------------------------------------------
n
o
an
------------
| 0 | SELECT STATEMENT | | 364 | 35308 | 17
(0)| 00:00:01 |
h s
a deฺ
|* 1 | TABLE ACCESS FULL| PROMOTIONS |
e )
s Gui
364 | 35308 | 17
r ฺ
mi dent
(0)| 00:00:01 |
p
g Stuฺ
--------------------------------------------------------------------
------------
d
b i
Predicate Informationa@(identified
t h is by operation id):
i a (ja use
---------------------------------------------------
A b to
i o
1 - filter("PROMO_ID">300)
nton
e A
J o s Statistics
----------------------------------------------------------
123 recursive calls
0 db block gets
88 consistent gets
3 physical reads
0 redo size
4 sorts (memory)
0 sorts (disk)
235 rows processed
SQL>
SQL> SELECT /*+ INDEX(promotions) */ *
2 FROM promotions
4 /
235 rows selected.

Elapsed: 00:00:00.01
Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
----------------------
| Id | Operation | Name | Rows | Bytes |
Cost (%CPU)| Time |
--------------------------------------------------------------------
----------------------
| 0 | SELECT STATEMENT | | 364 | 35308 |
353 (0)| 00:00:05 |
| 1 | TABLE ACCESS BY INDEX ROWID| PROMOTIONS | 364 | 35308 |
n s e
353 (0)| 00:00:05 |
i ce
|* 2 | INDEX RANGE SCAN | PROMO_PK | 364 | |
b l el
1 (0)| 00:00:01 |
--------------------------------------------------------------------
fe r a
----------------------
a n s
o n -tr
an
---------------------------------------------------
h s
a deฺ
2 - access("PROMO_ID">300)
e )
s Gui
r ฺ
i ent
p ฺ m d
Statistics
d g S t u
b
1 recursivea@ calls
t h is
----------------------------------------------------------
i
a blocksegets
0 (jdb
b i
243a t o u gets
consistent
nto redo size
J o s 585
17
0 sorts (memory)
0 sorts (disk)
235 rows processed
SQL>
SQL> set timing off
SQL>

SELECT *
FROM promotions_iot
a) The optimizer directly uses the index-organized structure, which is extremely
efficient in this case compared to the previous step.
SQL> @query14

SQL> set echo on
SQL>
SQL> set timing on
SQL>
SQL> SELECT *
2 FROM promotions_iot
4 /
235 rows selected.
Elapsed: 00:00:00.00
Execution Plan
n s e
----------------------------------------------------------
i ce
b l el
fer
-------------------------------------------------------------------- a
------------------
a n s
| Id | Operation | Name -tr
| Rows | Bytes | Cost
n
o
an
(%CPU)| Time |
--------------------------------------------------------------------
------------------
h s
a deฺ
e )
s Gui | 235 | 23500 | 2
r ฺ
mi dent
(0)| 00:00:01 |
p ฺ
|* 1 | INDEX RANGE SCAN| SYS_IOT_TOP_72170 |
g Stu
235 | 23500 | 2
(0)| 00:00:01 |
d
b
------------------ i a@ this
--------------------------------------------------------------------

b
Predicate
A to
Information
o
ni 1 - access("PROMO_ID">300)
---------------------------------------------------
n t o
e A
Jos Note
-----
- dynamic sampling used for this statement
Statistics
----------------------------------------------------------
5 recursive calls
0 db block gets
42 consistent gets
0 physical reads
116 redo size
0 sorts (memory)
0 sorts (disk)
235 rows processed
SQL>

SQL> set timing off
SQL>
44) Execute the iot_cleanup.sql script to clean up your environment.

SQL> @iot_cleanup
SQL> set echo on

SQL>
SQL> drop table promotions_iot purge;
Table dropped.
SQL>
SQL>
45) Case 14: Execute shc_setup.sql to set up your environment for this lab. n s e
i ce
SQL> @shc_setup
b l el
SQL> set echo on
fer a
SQL>
a n s
SQL>
n - tr
o
an
SQL> drop cluster bigemp_cluster including tables;
drop cluster bigemp_cluster including tables
* h s
a deฺ
ERROR at line 1:
e )
s Gui
r ฺ
ORA-00943: cluster does not exist
p ฺ mi dent
d g Stu
SQL>
b i a@ t h is
a
2 (deptno(jnumber,
SQL> CREATE CLUSTER
e number sort)
bigemp_cluster
ssal
b i a u
toHASH IS deptno SIZE 50
o A
3 HASHKEYS 10000
t o ni5 tablespace users;

4 single table
An
e
Jos
Cluster created.
SQL>
SQL> create table bigemp_fact (
2 empno number primary key, sal number sort, job varchar2(12) not
null,
3 deptno number not null, hiredate date not null)
4 CLUSTER bigemp_cluster (deptno, sal);
Table created.
SQL>
SQL>
SQL> begin
2 for i in 1..1400000 loop
3 insert into bigemp_fact values(i,i,'J1',10,sysdate);
4 end loop;
5 commit;
6 end;
7 /

SQL>
SQL> begin
2 for i in 1..1400000 loop
3 insert into bigemp_fact values(1400000+i,i,'J1',20,sysdate);
4 end loop;
5 commit;
6 end;
7 /
SQL>
SQL>
n s e
ice
b l el
SQL>
fe r a
a n s
46) Execute the query15.sql script. What do you observe?
o n -tr
a) Because you may have a lot of memory on your system, n the script first reduces
a optimizer
the amount of memory available to your session.
h a sThe
e ฺ decides to use the
s) costG
cluster access path to retrieve the data. The ui d
is minimal.
r ฺ e
SQL> @query15
p ฺ mi dent
SQL> set echo on
d g Stu
SQL>
SQL> set linesizeia
@ his
a on se t
b 200
SQL> set timing( j
a to utraceonly
SQL> set iautotrace
SQL> A
b
o n io alter session set workarea_size_policy=manual;
SQL>
nt
e A Session altered.
J o s
Elapsed: 00:00:00.02
SQL> alter session set sort_area_size=50000;
Session altered.
Elapsed: 00:00:00.01
SQL>
System altered.
Elapsed: 00:00:00.16
System altered.
Elapsed: 00:00:06.69
SQL>
SQL>

SQL> select * from bigemp_fact where deptno=10;
Elapsed: 00:00:07.21
Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
-------------
(%CPU)| Time |
--------------------------------------------------------------------
-------------
n s e
| 0 | SELECT STATEMENT | | 1400K| 32M| 1
ice
(0)| 00:00:01 |
b l el
|* 1 | TABLE ACCESS HASH| BIGEMP_FACT | 1400K|
(0)| 00:00:01 |
32M| 1
fe ra
--------------------------------------------------------------------
a n s
-------------
o n -tr
Predicate Information (identified by operation
s a nid):
h a deฺ
---------------------------------------------------
)
r ฺ e s Gui
mi dent
1 - access("DEPTNO"=10)
p ฺ
g Stu
d
Statistics
b i a@ this
----------------------------------------------------------
a
689 (jrecursive segets
calls
b i a
0 db
to u
block
i o A99451 consistent gets
to n 5999 physical reads
A n 116 redo size
o s e 33605911 bytes sent via SQL*Net to client

J 1027083
93335
10 sorts (memory)
0 sorts (disk)
SQL>
SQL> set timing off
SQL>

a) Again, the script first ensures that the amount of memory available to your session
is reduced. Then the script executes the same query, but asks for ordering the
result based on the sorted sal column. The optimizer can still use the cluster
access path without sorting the data. The cost is still minimal.
SQL> @query16
SQL> set echo on

SQL>
SQL> set timing on
SQL>
SQL> alter session set workarea_size_policy=manual;
Session altered.
Elapsed: 00:00:00.00
Session altered.
Elapsed: 00:00:00.00
SQL>
n s e
i ce
b l el
System altered.
fe r a
Elapsed: 00:00:00.10
a n s
o n -tr
System altered.
s an
) h a deฺ
Elapsed: 00:00:00.04
r ฺ e s Gui
miwherededeptno=10
nt
SQL>
SQL> select * from bigemp_fact
g p ฺ u
order by sal;
d S t
@ s
thi
a b ia
j e
ia ( to us
Elapsed: 00:00:07.41
b
i o A Plan
Execution
n
ntoPlan hash value: 865757019
----------------------------------------------------------
e A
J o s --------------------------------------------------------------------
-------------
(%CPU)| Time |
--------------------------------------------------------------------
-------------
(0)| 00:00:01 |
|* 1 | TABLE ACCESS HASH| BIGEMP_FACT | 1400K| 32M| 1
(0)| 00:00:01 |
--------------------------------------------------------------------
-------------

---------------------------------------------------
Statistics

----------------------------------------------------------
10 db block gets
6012 physical reads
0 redo size

12 sorts (memory)
0 sorts (disk)
SQL>
SQL> set timing off
n s e
SQL>
ice
48) Execute the query17.sql script. What do you observe? b l el
fe ra
a n s
n - tr
is reduced. Then the script executes the same query, but asks to order the result
o
based on the sorted sal column in the descending order. The optimizer can still
an
use the cluster access path without sorting the data. The cost is still minimal.
s
) h a deฺ
SQL> @query17
r ฺ e s Gui
mi dent
SQL> set echo on
SQL>
SQL> set linesize 200 dg p ฺ
SQL> set timing on @ S tu
b i atraceonly
t h is
(ja use
SQL> set autotrace
SQL>
i a
b sessionto set workarea_size_policy=manual;
A
SQL> alter
o
ni
toSession altered.
A n
o s e Elapsed: 00:00:00.00
J SQL> alter session set sort_area_size=50000;
Session altered.
Elapsed: 00:00:00.00
SQL>
System altered.
Elapsed: 00:00:00.09
System altered.
Elapsed: 00:00:00.12
SQL>
SQL> select * from bigemp_fact where deptno=10 order by sal desc;

Elapsed: 00:00:07.35
Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
-------------
(%CPU)| Time |
--------------------------------------------------------------------
-------------
(0)| 00:00:01 |
|* 1 | TABLE ACCESS HASH| BIGEMP_FACT | 1400K| 32M| 1
n s e
(0)| 00:00:01 |
ice
--------------------------------------------------------------------
b l el
-------------
fe r a
a n s
---------------------------------------------------
o n -tr
s an
) h a deฺ
r ฺ e s Gui
Statistics
ฺ m i ent
----------------------------------------------------------
d
1090 recursive calls g p t u d
10 db block
a@gets is S
a b i
99509 consistent t h
gets
6005 (jphysical
i a u sereads
A b bytes
0
to sent via SQL*Net to client
redo size
i o
33605911
nton 1027083 bytes received via SQL*Net from client
e A 93335 SQL*Net roundtrips to/from client
J o s 12
0
sorts (memory)
sorts (disk)
SQL>
SQL> set timing off
SQL>

is reduced. Then the script executes the same query but this time asks to order the
result based on the nonsorted empno column. The optimizer can still make use of
the cluster access path, but must sort the data making the cost of the query higher.
SQL> @query18
SQL> set echo on
SQL>
SQL> set timing on

SQL>
Session altered.
Elapsed: 00:00:00.00
Session altered.
Elapsed: 00:00:00.00
SQL>
System altered.
n s e
i ce
Elapsed: 00:00:00.10
b l el
fer a
System altered.
a n s
o n -tr
an
Elapsed: 00:00:00.04
SQL>
h s
a deฺ
SQL> select * from bigemp_fact where deptno=10 order by empno;
e )
s Gui
1400000 rows selected. r ฺ
i ent
p ฺ m
Elapsed: 00:00:10.01
d g Stud
Execution Plan bia
@ his
j a e t
( s
----------------------------------------------------------
u
iavalue:to1775608660
b
Plan hash
n i oA
nto----------------------
--------------------------------------------------------------------
e A
J o s | Id | Operation
Cost (%CPU)| Time |
| Name | Rows | Bytes |TempSpc|
--------------------------------------------------------------------
----------------------
| 0 | SELECT STATEMENT | | 1400K| 32M| |
47728 (1)| 00:09:33 |
| 1 | SORT ORDER BY | | 1400K| 32M| 107M|
47728 (1)| 00:09:33 |
|* 2 | TABLE ACCESS HASH| BIGEMP_FACT | 1400K| 32M| |
1 (0)| 00:00:01 |
--------------------------------------------------------------------
----------------------

---------------------------------------------------
Statistics
----------------------------------------------------------

12 db block gets
0 redo size

12 sorts (memory)
1 sorts (disk)
SQL>
SQL> set timing off
SQL>
n s e
ice
b l el
r a
fe
n s
is reduced. Then the script executes the same query, but this time asks to order the
a
- tr
result based on the sal, deptno key. The optimizer can still make use of the
n
o
an
cluster access path, but must sort the data making the cost of the query higher.
s
SQL> @query19
) h a deฺ
SQL> set echo on
r ฺ e s Gui
mi dent
SQL>
p ฺ
g Stu
SQL> set timing on d
b i a@ this
(ja usetseworkarea_size_policy=manual;
SQL>
i a
SQL> alter session
baltered.to
i o A
Session
nton
e A Elapsed: 00:00:00.00
J o s SQL> alter session set sort_area_size=50000;
Session altered.
Elapsed: 00:00:00.00
SQL>
System altered.
Elapsed: 00:00:00.10
System altered.
Elapsed: 00:00:00.09
SQL>
SQL> select * from bigemp_fact where deptno=10 order by sal,empno;

Elapsed: 00:00:09.25
Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
----------------------
| Id | Operation | Name | Rows | Bytes |TempSpc|
Cost (%CPU)| Time |
--------------------------------------------------------------------
----------------------
| 0 | SELECT STATEMENT | | 1400K| 32M| |
47728 (1)| 00:09:33 |
| 1 | SORT ORDER BY | | 1400K| 32M| 107M|
47728 (1)| 00:09:33 |
n s e
|* 2 | TABLE ACCESS HASH| BIGEMP_FACT | 1400K| 32M| |
i ce
1 (0)| 00:00:01 |
b l el
--------------------------------------------------------------------
----------------------
fer a
a n s
o n -tr
an
---------------------------------------------------
h s
a deฺ
e )
s Gui
r ฺ
i ent
Statistics
p ฺ m d
d g t u
----------------------------------------------------------
S
12 db b i a@gets
1139 recursive
block
calls
t h is
a sereads
6178 (jconsistent gets
i a
b 0 redotosize
12238 u
physical
i o A
nton 33605911 bytes sent via SQL*Net to client
e A 1027083 bytes received via SQL*Net from client
J o s 93335
12
sorts (memory)
1 sorts (disk)
SQL>
SQL> set timing off
SQL>
51) Execute the shc_cleanup.sql script to clean up your environment.

SQL> @shc_cleanup
SQL> set echo on
SQL>
SQL> drop cluster bigemp_cluster including tables;
Cluster dropped.
SQL>
SQL>

52) Case 15: Execute the nic_setup.sql script to set up your environment for this
case.
SQL> @nic_setup
SQL> set echo on
SQL>
SQL> drop cluster emp_dept including tables;

drop cluster emp_dept including tables
*
ERROR at line 1:
SQL>
SQL> drop table emp purge;
drop table emp purge
n s e
*
i ce
ERROR at line 1:
b l el
fer a
a n s
SQL> drop table dept purge;
n - tr
o
an
drop table dept purge
*
ERROR at line 1:
h s
a deฺ
e
ORA-00942: table or view does not exist )
s Gui
r ฺ
p ฺ mi dent
SQL> d g Stu
SQL>
i @( his
aemp
SQL> CREATE TABLE
j b
a NUMBER(7)e t
2 ( u s
ia toVARCHAR2(15)
empno ,
3 b
4o A job
ename NOT NULL,
o n i VARCHAR2(9) ,
nt 6 hiredate DATE
5 mgr NUMBER(7) ,
e A ,
Jos
7 sal NUMBER(7) ,
8 comm NUMBER(7) ,
9 deptno NUMBER(3)
10 );
Table created.
SQL>
SQL> CREATE TABLE dept (
2 deptno NUMBER(3) ,
3 dname VARCHAR2(14),
4 loc VARCHAR2(14),
5 c VARCHAR2(500)
6 );
Table created.
SQL>
SQL> CREATE INDEX emp_index
2 ON emp(deptno)
3 TABLESPACE users

4 STORAGE (INITIAL 50K
5 NEXT 50K
6 MINEXTENTS 2
7 MAXEXTENTS 10
8 PCTINCREASE 33);
Index created.
SQL>
SQL> CREATE INDEX dept_index
2 ON dept(deptno)
3 TABLESPACE users
5 NEXT 50K
6 MINEXTENTS 2
7 MAXEXTENTS 10
n s e
8 PCTINCREASE 33);
i ce
b l el
Index created.
fer a
SQL>
a n s
SQL>
n - tr
o
an
SQL> begin
2 for i in 1..999 loop
3 insert into dept values
h s
a deฺ
)
s Gui
(i,'D'||i,'L'||i,dbms_random.string('u',500));
e
r ฺ
mi dent
4 end loop;
5 commit;
p ฺ
g Stu
6 end;
d
7 /
b i a@ this
i a (ja successfully
PL/SQL procedure
u se completed.
SQL> A
b to
n i o
t oSQL>
n SQL> begin
e A
Jos
3 insert into emp values
(i,dbms_random.string('u',15),dbms_random.string('u',9),i,sysdate,i,
i,mod(i,999));
4 end loop;
5 commit;
6 end;
7 /
SQL>
SQL>
53) Execute the nic_query.sql script. What do you observe?

a) The script first ensures that the amount of memory available to your session is
reduced. Then the script executes a join between the EMP and DEPT tables. The
optimizer is able to make use of the index to resolve the join.
SQL> @nic_query
SQL> set echo on
SQL>
SQL> set timing on
SQL>
Session altered.
Elapsed: 00:00:00.00
n s e
i ce
el
a b l
Session altered.
fe r
a n s
-tr
Elapsed: 00:00:00.00
SQL> alter session set hash_area_size=5000;
o n
Session altered. s an
) h a deฺ
Elapsed: 00:00:00.00
r ฺ e s Gui
SQL>
SQL> p ฺ mi dent
d g where
SQL> select * from emp,dept
S tuemp.deptno=dept.deptno and
emp.deptno > 800; a@ is
b i t h
(ja use
99000 rowsaselected.
i
A b to
i o
ton
Elapsed: 00:00:02.88
A n Execution Plan
o s e ----------------------------------------------------------
J Plan hash value: 128236434
--------------------------------------------------------------------
-------------------------------
|TempSpc| Cost (%CPU)| Time |
--------------------------------------------------------------------
-------------------------------
| 0 | SELECT STATEMENT | | 19780 | 10M|
| 3449 (1)| 00:00:42 |
|* 1 | HASH JOIN | | 19780 | 10M|
104K| 3449 (1)| 00:00:42 |
| 2 | TABLE ACCESS BY INDEX ROWID| DEPT | 199 | 99K|
| 18 (0)| 00:00:01 |
|* 3 | INDEX RANGE SCAN | DEPT_INDEX | 199 | |
| 2 (0)| 00:00:01 |
|* 4 | TABLE ACCESS FULL | EMP | 99198 | 5521K|
| 1207 (2)| 00:00:15 |
--------------------------------------------------------------------
-------------------------------

---------------------------------------------------
1 - access("EMP"."DEPTNO"="DEPT"."DEPTNO")
3 - access("DEPT"."DEPTNO">800)
4 - filter("EMP"."DEPTNO">800)
Statistics
----------------------------------------------------------
9 recursive calls
0 db block gets
0 redo size
n s e
i ce
b l el
0 sorts (memory)
fer a
0 sorts (disk)
a n s
o n -tr
SQL>
s an
) h a deฺ
SQL> set timing off;
r ฺ e s Gui
mi ofdthe nt
SQL>
p ฺ e
g Stu previous query? Implement your
54) How would you enhance the performance
d
solution.
b i a@ this
i a (ja to store
a) Create a cluster
u sethe two tables.
A b
SQL> @ic_setup to
o
ni set echo on
SQL>
t o
An SQL> drop table emp purge;
SQL>
e
Jos Table dropped.
Table dropped.
SQL>
drop cluster emp_dept including tables
*
ERROR at line 1:
SQL>
SQL> CREATE CLUSTER emp_dept (deptno NUMBER(3))
2 SIZE 600
3 TABLESPACE users
5 NEXT 300K

6 MINEXTENTS 2
7 PCTINCREASE 33);
Cluster created.
SQL>
SQL> CREATE TABLE emp (

2 empno NUMBER(7) ,
3 ename VARCHAR2(15) NOT NULL,
4 job VARCHAR2(9) ,
5 mgr NUMBER(7) ,
6 hiredate DATE ,
7 sal NUMBER(7) ,
8 comm NUMBER(7) ,
9 deptno NUMBER(3))
10 CLUSTER emp_dept (deptno);
n s e
i ce
Table created.
b l el
SQL>
fer a
SQL> CREATE TABLE dept (
a n s
2 deptno NUMBER(3) ,
n -tr
o
an
3 dname VARCHAR2(14),
4 loc VARCHAR2(14),
5 c VARCHAR2(500))
h s
a deฺ
6 CLUSTER emp_dept (deptno);
e )
s Gui
r ฺ
Table created.
p ฺ mi dent
d g Stu
SQL>
i @ his
aemp_dept_index
SQL> CREATE INDEX
j b
a emp_depte t
2 (
ON CLUSTER
ia to uusers s
3 bTABLESPACE
4 A STORAGE (INITIAL 50K
n i o
o
nt 6
5 NEXT 50K
e A MINEXTENTS 2
Jos
7 MAXEXTENTS 10
8 PCTINCREASE 33);
Index created.
SQL>
SQL> begin
3 insert into dept values
(i,'D'||i,'L'||i,dbms_random.string('u',500));
4 end loop;
5 commit;
6 end;
7 /
SQL>
SQL>
SQL> begin

3 insert into emp values
(i,dbms_random.string('u',15),dbms_random.string('u',9),i,sysdate,i,
i,mod(i,999));
4 end loop;
5 commit;
6 end;
7 /
SQL>
SQL>
n s e
i ce
55) Execute the following query to confirm the performance enhancement of the previous
query: b l el
select * fe r a
from emp,dept
a n s
where emp.deptno=dept.deptno and emp.deptno > 800;
o n -tr
s an
h
a) The optimizer is able to use the cluster access
) i eฺ the query execute
apath thatdmakes
faster.
i r ฺes nt Gu
SQL> @ic_query
g p ฺm ude
SQL> set echo on d S t
@ s
SQL>
j a b ia
e thi
SQL> set timing
i a (
SQL> set autotrace
on
u s
traceonly
b
A linesizet 200
SQL> set o
i o
n select * from emp,dept where emp.deptno=dept.deptno and
SQL>
ntoSQL>
e A emp.deptno > 800;
J o s
Elapsed: 00:00:01.20
Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
--------------------
| Id | Operation | Name | Rows | Bytes |
Cost (%CPU)| Time |
--------------------------------------------------------------------
--------------------
| 0 | SELECT STATEMENT | | 19780 | 10M|
11515 (1)| 00:02:19 |
| 1 | NESTED LOOPS | | 19780 | 10M|
11515 (1)| 00:02:19 |
| 2 | TABLE ACCESS CLUSTER| DEPT | 199 | 99K|
167 (0)| 00:00:03 |

|* 3 | INDEX RANGE SCAN | EMP_DEPT_INDEX | 1 | |
2 (0)| 00:00:01 |
|* 4 | TABLE ACCESS CLUSTER| EMP | 99 | 5643 |
57 (0)| 00:00:01 |
--------------------------------------------------------------------
--------------------

---------------------------------------------------
3 - access("DEPT"."DEPTNO">800)
4 - filter("EMP"."DEPTNO">800 AND "EMP"."DEPTNO"="DEPT"."DEPTNO")
Statistics
----------------------------------------------------------
n s e
1 recursive calls
i ce
0 db block gets
b l el
0 physical reads
fer a
0 redo size
a n s
o n -tr
an
0 sorts (memory)
h s
a deฺ
0 sorts (disk)
e )
s Gui
i ent
p ฺ m d
SQL>
d g S t u
SQL> set timing b i
offa@ this
SQL>
i a (ja use
A
56) Execute
b to
the ic_cleanup.sql script to clean up your environment for this case.
n i o
toSQL> @ic_cleanup
e An SQL> set echo on
Jos SQL>
Cluster dropped.
SQL>
*
ERROR at line 1:

drop table dept purge
*
ERROR at line 1:
SQL>
SQL>

57) Case 16: Execute the iss_setup.sql script to set up your environment for this
lab.
SQL> @iss_setup
SQL> set echo on
SQL>
SQL> create table t(c number, d number);
Table created.
SQL>
SQL> begin
3 insert into t values(1,i);
4 end loop;
5 end;
n s e
6 /
i ce
b l el
fer a
SQL>
a n s
SQL> create index it on t(c,d);
o n -tr
Index created.
s an
) h a deฺ
SQL>
r ฺ e s Gui
p ฺ mi dent
select count(*) from t where
d g d=1;S tu
b i a@ this
a (jais notuusing
a) The optimizer
i se the index and does a full table scan.
A b to
o
ni set echo on
SQL> @query20
t oSQL>
e An SQL>
Jos
SQL>
SQL> set timing on
SQL> set autotrace on
SQL>
SQL> select count(*) from t where d=1;
COUNT(*)
----------
1
Elapsed: 00:00:00.01
Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
-------
Time |

--------------------------------------------------------------------
-------
00:00:01 |
| 1 | SORT AGGREGATE | | 1 | 13 | |
|
|* 2 | TABLE ACCESS FULL| T | 1 | 13 | 7 (0)|

00:00:01 |
--------------------------------------------------------------------
-------

---------------------------------------------------
2 - filter("D"=1)
n s e
Note
i ce
-----
b l el
fe ra
a n s
Statistics
n - tr
o
an
----------------------------------------------------------
5 recursive calls
0 db block gets
h s
a deฺ
48 consistent gets
e )
s Gui
r ฺ
mi dent
0 physical reads
0 redo size
p ฺ
g Stu
d
b i a@ this
i a (ja use
0 sorts (memory)
A b to
0 sorts (disk)
o n io 1 rows processed
nt SQL>
e A
J o s SQL> set timing off
SQL>
59) How would you improve the performance of a query, such as the one in the previous
step? Implement your solution.
a) Make sure you gather correctly the statistics for your table so that the index skip
scan can be used.
SQL> @iss_gather_stats
SQL> set echo on
SQL>
SQL> execute dbms_stats.gather_table_stats('SH','T',cascade=>TRUE);
SQL>

select count(*) from t where d=1;

a) The optimizer now uses the index to perform an index skip scan.
SQL> @query20
SQL> set echo on
SQL>

SQL>
SQL> set timing on
SQL>
SQL> select count(*) from t where d=1;
COUNT(*)
----------
n s e
ce
1
eli
Elapsed: 00:00:00.00
a b l
fer
Execution Plan
a n s
o n -tr
----------------------------------------------------------
s an
) h a deฺ
--------------------------------------------------------------------
ui | Cost (%CPU)| Time

-----
| Id | Operation | Name | e
r ฺ s | GBytes
Rows
|
p ฺ mi dent
g Stu
--------------------------------------------------------------------
d
a@ t|his | 1 | 4 | 2 (0)|
-----
|
b i
00:00:01 |
a (jaAGGREGATE
1 | iSORT u se |
|
| b t o | 1 |
A INDEX SKIP SCAN| IT | 1 | 4 | 2 (0)|
4 | |
|*io2 |
n
to00:00:01 |
A n
o s e --------------------------------------------------------------------
-----
J
---------------------------------------------------
2 - access("D"=1)
filter("D"=1)
Statistics
----------------------------------------------------------
0 recursive calls
0 db block gets
23 consistent gets
0 physical reads
0 redo size
0 sorts (memory)

0 sorts (disk)
1 rows processed
SQL>
SQL> set timing off
SQL>
61) Compare the result of executing the previous query with the result you obtain when
you execute the following query:
select /*+ INDEX_FFS(t it) */ count(*) from t where d=1;
a) The optimizer uses a fast full index scan, but this is not better than the index skip
scan.
n s e
ice
el
SQL> @query21
SQL> set echo on
a b l
SQL>
fe r
a n s
SQL>
SQL> set timing on
o n -tr
s an
etฺ where d=1;
SQL>
) h a fromd
SQL> select /*+ INDEX_FFS(t it) */ count(*)
i
i r ฺes nt Gu
COUNT(*)
----------
g p ฺm ude
1 d S t
@ s
j a b ia
e thi
i a (
Elapsed: 00:00:00.00
u s
b
A Plan t
Execution o
i o
n hash value: 273610729
----------------------------------------------------------
ntoPlan
e A
J o s --------------------------------------------------------------------
----------
Time |
--------------------------------------------------------------------
----------
00:00:01 |
| 1 | SORT AGGREGATE | | 1 | 4 | |
|
|* 2 | INDEX FAST FULL SCAN| IT | 1 | 4 | 9 (0)|
00:00:01 |
--------------------------------------------------------------------
----------

---------------------------------------------------
2 - filter("D"=1)

Statistics
----------------------------------------------------------
1 recursive calls
0 db block gets
31 consistent gets
0 physical reads
0 redo size
0 sorts (memory)
0 sorts (disk)
1 rows processed
SQL>
SQL> set timing off
n s e
i ce
SQL>
b l el
r a
62) Execute the iss_cleanup.sql script to clean up your environment for this case.
fe
a n s
SQL> @iss_cleanup
SQL> set echo on
o n -tr
SQL>
s an
SQL> drop table t purge;
) h a deฺ
Table dropped.
r ฺ e s Gui
p ฺ mi dent
SQL>
d g Stu
a@session,
SQL>
b i t h is
j a
63) Exit from your(SQL*Plus
s e and execute the ap_cleanup.sh script to clean
u
bia tofor this lab.
up your environment
A
io exit;
o n
SQL>
t Disconnected from Oracle Database 11g Enterprise Edition Release
e An 11.1.0.6.0 - Production
Jos With the Partitioning, Oracle Label Security, OLAP, Data Mining
[oracle@edrsr33p1-orcl Access_Paths]$ ./ap_cleanup.sh
Connected to:
Production
SQL>
SQL> revoke dba from sh;
Revoke succeeded.

SQL>
SQL> @sh_main sh example temp oracle
/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/
/home/oracle/ v3
SQL> Rem
SQL> Rem $Header: sh_main.sql 06-mar-2008.15:00:45 cbauwens Exp $
SQL> Rem
....
<<<<< CHANNELS DIMENSION >>>>>

Dimension - display name, description and plural name
Level - display name and description
Hierarchy - display name and description
- default calculation hierarchy
- default display hierarchy
n s e
Level Attributes - name, display name, description
i ce
Drop dimension attributes prior to re-creation
b l el
No attribute to drop
Create dimension attributes and add their level attributes
fer a
- Long Description created
a n s
- Short Description created
n -tr
o
an
Classify entity descriptor use
- Long Description
- Short Description
h s
a deฺ
-
e )
s Gui
r ฺ
mi dent
<<<<< FINAL PROCESSING >>>>>
p
g Sฺ
- Changes have been committed
d tu
b i a@ this
i a (ja use
b
Commit complete.
A to
ni o
ntogathering
e A statistics ...
J o s PL/SQL procedure successfully completed.

[oracle@edrsr33p1-orcl Access_Paths]$

Practice 4-2: Using the Result Cache
In this practice, you explore the various possibilities of caching query results in the
System Global Area (SGA). Perform the following steps to understand the use of Query
Result Cache.
1) Change to the $HOME/solutions/Query_Result_Cache directory and
execute the result_cache_setup.sh script.

[oracle@edrsr33p1-orcl ~]$ cd $HOME/solutions/Query_Result_Cache
[oracle@edrsr33p1-orcl Query_Result_Cache]$ ./result_cache_setup.sh
SQL*Plus: Release 11.1.0.6.0 - Production on Tue Mar 25 14:46:14

2008

n s e
i ce
b l el
Connected to:
fe ra
Production
a n s
-
n tr
n o
s a
SQL> SQL> SQL> SQL> drop user qrc cascade
) h a deฺ
*
r ฺ e s Gui
miexistdent
ERROR at line 1:
p ฺ
ORA-01918: user 'QRC' does not
g Stu
d
@ his
SQL> SQL> 2 bi3a t
( j a s e
ia to u
User created.
b
i o Asucceeded.
SQL> SQL>
n
nto
Grant
e A
Jos
SQL> SQL>
SQL> SQL> drop table cachejfv purge

*
ERROR at line 1:
SQL> SQL>
Table created.
SQL> SQL>
1 row created.
SQL>
1 row created.
SQL>
2 rows created.

Practice 4-2: Using the Result Cache (continued)
SQL>
4 rows created.
SQL>
8 rows created.
SQL>
16 rows created.
SQL>
32 rows created.
SQL>
64 rows created.
n s e
SQL>
i ce
128 rows created.
b l el
SQL>
fer a
256 rows created.
a n s
n -tr
o
an
SQL>
512 rows created.
h s
a deฺ
SQL>
e )
s Gui
1024 rows created. r ฺ
i ent
p ฺ m d
SQL>
2048 rows created. @ d g S t u
b i a t h is
SQL>
i a (ja use
b
4096 rows
A to
created.
n i o
ntoSQL>
e A 8192 rows created.
J o s SQL>
16384 rows created.
SQL>
32768 rows created.
SQL>
65536 rows created.
SQL>
SQL>
SQL>
SQL>

SQL> SQL>
1 row created.
SQL> SQL>
Commit complete.
SQL> SQL>
System altered.

[oracle@edrsr33p1-orcl Query_Result_Cache]$
n s e
i ce
--------------------------------------------------------------
b l el
fer a
#!/bin/bash
a n s
o n -tr
an
cd /home/oracle/solutions/Query_Result_Cache
h s
a deฺ
e )
s Gui
r ฺ
i ent
p ฺ m d
export
d g S t u
b i a@ this
l/bin:/usr/X11R6/bin:/usr/java/jdk1.5.0_11/bin
sqlplus b/ia
(ja u<<FIN!
se
A to
as sysdba
o
ni echo on
ntoset
e A
J o s drop user qrc cascade;
create user qrc identified by qrc

default tablespace users
temporary tablespace temp;
grant connect, resource, dba to qrc;
connect qrc/qrc
exec dbms_result_cache.flush;
drop table cachejfv purge;
create table cachejfv(c varchar2(500)) tablespace users;
insert into cachejfv

values('aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa');
insert into cachejfv select * from cachejfv;


n s e
i ce
b l el
fer a
insert into cachejfv values('b');
a n s
o n -tr
an
commit;

h s
a deฺ
e )
s Gui
FIN! r ฺ
i ent
p ฺ m
2) Open a terminal window, invoke
d g SQL*Plus,
S t udand connect as the qrc user. From now
on, do not disconnect from
i a @this hsession.
i s Determine the current content of the query
j a b
cache using the following t
statement:
e
i a ( u s
select type,status,name,object_no,row_count,row_size_avg
from A b to
v$result_cache_objects order by 1;
i o
n do you observe?
n toWhat
e A a) Use the check_result_cache.sql script. Right now, the query cache
Jos should be empty.
[oracle@edrsr33p1-orcl Query_Result_Cache]$ sqlplus qrc/qrc

2008
Connected to:
Production
SQL> @check_result_cache
SQL>
SQL> select type,status,name,object_no,row_count,row_size_avg from

no rows selected
SQL>
SQL>
--------------------------------------------------------------
set echo on
select type,status,name,object_no,row_count,row_size_avg from

3) Set timing on and execute the query as follows. You can use the query1.sql
n s e
script. Note the time that it takes to execute.
i ce
select /*+ result_cache q_name(Q1) */ count(*) b l el
from cachejfv c1,cachejfv c2,cachejfv c3,cachejfv
fer a
c4,cachejfv c5,cachejfv c6, cachejfv c7
a n s
n -tr
where c1.c='b' and c2.c='b' and c3.c='b' and c4.c='b' and
o
an
c5.c='b' and c6.c='b' and c7.c='b';
SQL> set timing on
h s
a deฺ
SQL> @query1
e )
s Gui
r ฺ
COUNT(*)
p ฺ mi dent
----------
1 d g Stu
b i a@ this
SQL>
i a (ja use
Elapsed: 00:00:02.46
A b to
i o
ton
--------------------------------------------------------------
A n select /*+ result_cache q_name(Q1) */ count(*)

o s e from cachejfv c1,cachejfv c2,cachejfv c3,cachejfv c4,cachejfv
J c5,cachejfv c6, cachejfv c7
where c1.c='b' and c2.c='b' and c3.c='b' and c4.c='b' and c5.c='b'
and c6.c='b' and c7.c='b';
4) Determine the execution plan of the previous query by using the

explain_query1.sql script. What do you observe?
a) Because of the result_cache hint, the result of the query is computed using the
result cache.
SQL> @explain_query1
SQL> set echo on
SQL>
SQL> explain plan for
2 select /*+ result_cache q_name(Q1) */ count(*)
3 from cachejfv c1,cachejfv c2,cachejfv c3,cachejfv c4,cachejfv
c5,cachejfv c6, cachejfv c7
4 where c1.c='b' and c2.c='b' and c3.c='b' and c4.c='b' and

Explained.
Elapsed: 00:00:00.05
SQL>
SQL> set pagesize 200

SQL>
SQL> select plan_table_output from
table(dbms_xplan.display('plan_table',null,'ALL'));
PLAN_TABLE_OUTPUT
--------------------------------------------------------------------
--------------------------------------------------------------------
--------------------------------------------
n s e
i ce
--------------------------------------------------------------------
b l el
--------------------------------------
fe
|r a
a n s
- tr
--------------------------------------------------------------------
n
o
an
--------------------------------------
1 | 1764 | 7316P (1)|999:59:59 |
h s
a deฺ
| 1 | RESULT CACHE )
s Gui
| b7rh5vw33py4ug4n8xaav6525g |
e
r ฺ
mi dent
| | | |
| 2 | SORT AGGREGATE
p ฺ
g Stu
| |
1 | 1764 | |
d |
| 3 |
294P| b i a@ this
MERGE JOIN CARTESIAN
15E| 7316P (1)|999:59:59 |
| |
| 4 |
i a (ja use
MERGE JOIN CARTESIAN | |
b
941T| 1264P|
A to
23P (1)|999:59:59 |
o
|
n io
5 |
3007G| 3446T|
MERGE JOIN CARTESIAN |
74T (1)|999:59:59 |
|
nt
e A | 6 | MERGE JOIN CARTESIAN | |
J o s 9606M| 9018G|
| 7 |
238G (1)|999:59:59 |
MERGE JOIN CARTESIAN | |
30M| 21G| 761M (1)|999:59:59 |
| 8 | MERGE JOIN CARTESIAN| |
98011 | 47M| 2432K (1)| 08:06:26 |
|* 9 | TABLE ACCESS FULL | CACHEJFV |
313 | 78876 | 7748 (1)| 00:01:33 |
| 10 | BUFFER SORT | |
313 | 78876 | 2424K (1)| 08:04:53 |
313 | 78876 | 7746 (1)| 00:01:33 |
313 | 78876 | 761M (1)|999:59:59 |
313 | 78876 | 7746 (1)| 00:01:33 |
313 | 78876 | 238G (1)|999:59:59 |
313 | 78876 | 7746 (1)| 00:01:33 |
313 | 78876 | 74T (1)|999:59:59 |

313 | 78876 | 7746 (1)| 00:01:33 |
313 | 78876 | 23P (1)|999:59:59 |
313 | 78876 | 7746 (1)| 00:01:33 |
313 | 78876 | 7316P (1)|999:59:59 |
313 | 78876 | 7746 (1)| 00:01:33 |
--------------------------------------------------------------------
--------------------------------------
Query Block Name / Object Alias (identified by operation id):

-------------------------------------------------------------
n s e
1 - SEL$1
i ce
9 - SEL$1 / C7@SEL$1
b l el
11
13
-
-
SEL$1
SEL$1
/
/
C6@SEL$1
C5@SEL$1
fer a
15 - SEL$1 / C4@SEL$1
a n s
17 - SEL$1 / C3@SEL$1
n - tr
o
an
19 - SEL$1 / C2@SEL$1
21 - SEL$1 / C1@SEL$1
h s eฺ
a did):
Predicate Information (identified by )operation i
i r ฺes nt Gu
---------------------------------------------------
g p ฺm ude
d
9 - filter("C7"."C"='b')
S t
@ s
thi
11 - filter("C6"."C"='b')
b ia
13 - filter("C5"."C"='b')
a
j e
ia ( to us
15 - filter("C4"."C"='b')
b
17 - filter("C3"."C"='b')
n i o Afilter("C2"."C"='b')
19 -
nto 21 - filter("C1"."C"='b')
e A
J o s Column Projection Information (identified by operation id):
-----------------------------------------------------------
1 - COUNT(*)[22]
2 - (#keys=0) COUNT(*)[22]
3 - (#keys=0) "C7"."C"[VARCHAR2,500], "C6"."C"[VARCHAR2,500],
"C5"."C"[VARCHAR2,500],
"C4"."C"[VARCHAR2,500], "C3"."C"[VARCHAR2,500],
"C2"."C"[VARCHAR2,500], "C1"."C"[VARCHAR2,500]
"C5"."C"[VARCHAR2,500],
"C2"."C"[VARCHAR2,500]
"C5"."C"[VARCHAR2,500],
"C5"."C"[VARCHAR2,500],

8 - (#keys=0) "C7"."C"[VARCHAR2,500], "C6"."C"[VARCHAR2,500]
9 - "C7"."C"[VARCHAR2,500]
10 - (#keys=0) "C6"."C"[VARCHAR2,500]
11 - "C6"."C"[VARCHAR2,500]
12 - (#keys=0) "C5"."C"[VARCHAR2,500]
13 - "C5"."C"[VARCHAR2,500]
14 - (#keys=0) "C4"."C"[VARCHAR2,500]
15 - "C4"."C"[VARCHAR2,500]
16 - (#keys=0) "C3"."C"[VARCHAR2,500]
17 - "C3"."C"[VARCHAR2,500]
18 - (#keys=0) "C2"."C"[VARCHAR2,500]
19 - "C2"."C"[VARCHAR2,500]
20 - (#keys=0) "C1"."C"[VARCHAR2,500]
21 - "C1"."C"[VARCHAR2,500]
Result Cache Information (identified by operation id):

n s e
------------------------------------------------------
i ce
b l el
1 - column-count=1; dependencies=(QRC.CACHEJFV);
attributes=(single-row); parameters=(nls); name="select /*+
fer a
result_cache q_name(Q1) */ count(*)
a n s
o n -tr
from cachejfv c1,cachejfv c2,cachejfv c3,cachejfv c4,cachejfv
an
c5,cachejfv c6, cac"
h s
a deฺ
Note
e )
s Gui
r ฺ
mithisdstatement
nt
-----
ฺ
- dynamic sampling used for
p
g Stu e
89 rows selected. @d
b i a t h is
i a (ja use
Elapsed: 00:00:00.03
SQL>
A b to
ni o
nto--------------------------------------------------------------
e A
J o s set echo on
explain plan for

select /*+ result_cache q_name(Q1) */ count(*)
set linesize 180

set pagesize 200
select plan_table_output from

5) Determine the current content of the query cache by using the

check_result_cache.sql script. What do you observe?
a) You can now see the result of your query cached.

SQL> set echo on
SQL>
SQL> set long 2000
SQL>
TYPE STATUS NAME

OBJECT_NO ROW_COUNT
---------- --------- -----------------------------------------------
--------------------------------------------------------------------
------------- ---------- ----------
ROW_SIZE_AVG
------------
Dependency Published QRC.CACHEJFV
n s e
71474 0
i ce
0
b l el
fer a
Result
0 1
Published select /*+ result_cache q_name(Q1) */ count(*)
a n s
n -tr
from cachejfv c1,cachejfv c2,cachejfv
o
an
c3,cachejfv c4,cachejfv c5,cachejfv c6, cac
5
h s
a deฺ
e )
s Gui
r ฺ
mi dent
Elapsed: 00:00:00.00
SQL>
p ฺ
Elapsed: 00:00:00.00@
dg Stu
b i a t h is
(ja use
SQL>
b i a to
--------------------------------------------------------------
o A
ni echo on
toset
A n
o s e set long 2000
J select type,status,name,object_no,row_count,row_size_avg from
6) Flush the buffer cache of your instance and rerun the query executed in step 3. What
do you observe?
a) The execution time for the query is now almost instantaneous.
System altered.
Elapsed: 00:00:00.05
SQL> @query1
SQL> select /*+ result_cache q_name(Q1) */ count(*)

COUNT(*)
----------
1
Elapsed: 00:00:00.00
SQL>
--------------------------------------------------------------

n s e
i ce
b l el
7) Insert a new row into the CACHEJFV table using the following statement: fer a
insert into cachejfv values('c');
a n s
o n -tr
n
a invalidated.
a) The corresponding result cache entry is automatically
h s
a deฺ
SQL> insert into cachejfv values('c');
e )
s Gui
r ฺ
1 row created.
p ฺ mi dent
Elapsed: 00:00:00.00 d
g Stu
SQL> commit;
b i a@ this
i a (ja use
b
Commit complete.
A to
ni o
e A SQL> set echo on
J o s SQL>
SQL> set long 2000
SQL>
TYPE STATUS NAME

OBJECT_NO ROW_COUNT
---------- --------- -----------------------------------------------
--------------------------------------------------------------------
------------- ---------- ----------
ROW_SIZE_AVG
------------
71474 0
0
Result Invalid select /*+ result_cache q_name(Q1) */ count(*)

0 1

5
Elapsed: 00:00:00.00
SQL>
--------------------------------------------------------------
set echo on
set long 2000

n s e
ice
b l
8) Execute your query from step 3 and step 6 again and check the result cache. What do el
you observe? fe r a
a n s
entry has been added for the new result. o n -tr
a) Again, it takes some time to execute the query. The result cache shows that a new
s an
SQL> @query1
) h i d eฺ
a*/ count(*)
es t Gc3,cachejfv
u
SQL> select /*+ result_cache q_name(Q1)
i r ฺ
2 from cachejfv c1,cachejfv c2,cachejfv
c5,cachejfv c6, cachejfv c7 m n
c4,cachejfv
d g pฺ tuanddec3.c='b' and c4.c='b' and

3 where c1.c='b' and c2.c='b'
i a @ his S
COUNT(*) (ja
b et
b ia to us
----------
n i oA 1
e A SQL>
J o s SQL> @check_result_cache
SQL> set echo on
SQL>
SQL> set long 2000
SQL>
TYPE STATUS NAME

OBJECT_NO ROW_COUNT
---------- --------- -----------------------------------------------
--------------------------------------------------------------------
------------- ---------- ----------
ROW_SIZE_AVG
------------
71474 0
0

0 1
5
Result Published select /*+ result_cache q_name(Q1) */ count(*)

0 1
5
Elapsed: 00:00:00.00
SQL>
n s e
--------------------------------------------------------------
i ce
b l el
fe ra
a n s
o n -tr
an
h s
a deฺ
e )
s Gui
set echo on r ฺ
i ent
p ฺ m
set long 2000
d g Stud
b i a@ this
j a e
ia ( to us
b
n io A a detailed result cache memory report.
9)oGenerate
t
e An SQL> set serveroutput on
Jos SQL> EXEC DBMS_RESULT_CACHE.MEMORY_REPORT(detailed=>true);
R e s u l t C a c h e M e m o r y R e p o r t
[Parameters]
Block Size = 1K bytes
Maximum Cache Size = 2080K bytes (2080 blocks)
Maximum Result Size = 104K bytes (104 blocks)
[Memory]
Total Memory = 116008 bytes [0.035% of the Shared Pool]
... Fixed Memory = 5132 bytes [0.002% of the Shared Pool]
....... Cache Mgr = 108 bytes
....... Memory Mgr = 124 bytes
....... Bloom Fltr = 2K bytes
....... State Objs = 2852 bytes
... Dynamic Memory = 110876 bytes [0.033% of the Shared Pool]
....... Overhead = 78108 bytes
........... Hash Table = 32K bytes (4K buckets)
........... Chunk Ptrs = 12K bytes (3K slots)
........... Chunk Maps = 12K bytes
........... Miscellaneous = 20764 bytes
....... Cache Memory = 32K bytes (32 blocks)

........... Unused Memory = 29 blocks
........... Used Memory = 3 blocks
............... Dependencies = 1 blocks (1 count)
............... Results = 2 blocks
................... SQL = 1 blocks (1 count)
................... Invalid = 1 blocks (1 count)
Elapsed: 00:00:00.12
SQL>
10) Execute your query from step 3 and step 6 again. What do you observe?
a) The query again uses the result that was previously cached.
SQL> @query1
n s e
ice
el
a b l
fe r
a n s
-tr
o n
an
COUNT(*)
----------
1 h s
a deฺ
e )
s Gui
r ฺ
Elapsed: 00:00:00.01
SQL> p ฺ mi dent
d g Stu
i a@ this
--------------------------------------------------------------
b
i a (ja use q_name(Q1) */ count(*)
select /*+ result_cache
A b
from cachejfv o
tcachejfv
c1,cachejfv c2,cachejfv c3,cachejfv c4,cachejfv
o
ni c1.c='b' and c2.c='b' and c3.c='b' and c4.c='b' and c5.c='b'
c5,cachejfv c6, c7
ntowhere
e A and c6.c='b' and c7.c='b';
J o s
11) Ensure that you bypass the result cache before performing the next step.
SQL>
SQL> exec DBMS_RESULT_CACHE.BYPASS(bypass_mode=>true);
Elapsed: 00:00:00.01
SQL>
12) Execute your query again. What do you observe?

a) The query again takes longer to execute because it no longer uses the result cache.
SQL> @query1

COUNT(*)
----------
1
Elapsed: 00:00:02.34
SQL>
--------------------------------------------------------------

n s e
i ce
13) Ensure that you no longer bypass the result cache and check that your query uses it
again. b l el
fe ra
SQL> exec DBMS_RESULT_CACHE.BYPASS(bypass_mode=>false);
a n s
PL/SQL procedure successfully completed. o n -tr
s an
Elapsed: 00:00:00.01
) h a deฺ
SQL> @query1
r ฺ e ui
s */Gcount(*)
2 from cachejfv c1,cachejfv
ฺ m ic2,cachejfv
SQL> select /*+ result_cache q_name(Q1)
e nt c3,cachejfv c4,cachejfv
c5,cachejfv c6, cachejfv gc7p d
tuand c3.c='b' and c4.c='b' and
3 where c1.c='b' @ anddc2.c='b'
S
b i
c5.c='b' and c6.c='b' is
a andthc7.c='b';
i a (ja use
b
COUNT(*)
A to
o
----------
i
nton 1
e A
J o s Elapsed: 00:00:00.00
SQL>
SQL>
--------------------------------------------------------------

14) Execute the following query by using the query2.sql script:

select count(*) from cachejfv c1,cachejfv c2,cachejfv
c3,cachejfv c4,cachejfv c5,cachejfv c6, cachejfv c7 where
c1.c='b' and c2.c='b' and c3.c='b' and c4.c='b' and c5.c='b'

a) Although the query is the same as the one used in step 3, it is not recognized as
cached because it does not contain the hint. So its execution time is long again.
SQL> @query2

COUNT(*)
----------
1
Elapsed: 00:00:02.41
SQL>
n s e
ice
--------------------------------------------------------------
b l el
select count(*) fe ra
a n s
o n -tr
an
h s
a deฺ
e )
suse the G ui result without using hints?
15) How would you force the previous query r ฺ to cached
Use the force_query2.sqlp ฺ miand dthen
script e t that you successfully used the
nverify
cached result. Finally, undo d g change.
your S tu
b i a@ this
i a
SQL> set echo (jaon use
SQL> @force_query2
SQL>
A b to result_cache_mode
ni
SQL>o show parameter
t o
e An NAME TYPE VALUE
Jos
------------------------------------ ----------- -------------------
-----------
result_cache_mode string MANUAL
SQL>
TYPE STATUS NAME

OBJECT_NO ROW_COUNT
---------- --------- -----------------------------------------------
--------------------------------------------------------------------
------------- ---------- ----------
ROW_SIZE_AVG
------------
71474 0
0

0 1

5
Result Published select /*+ result_cache q_name(Q1) */ count(*)

0 1

5
Elapsed: 00:00:00.00
SQL>
SQL> alter session set result_cache_mode=force;
Session altered.
n s e
i ce
Elapsed: 00:00:00.04
b l el
SQL>
fer a
2 select count(*)
a n s
-tr
n
o
an
h s
a deฺ
e )
s Gui
Explained. r ฺ
i ent
p ฺ m d
Elapsed: 00:00:00.05
d g S t u
SQL>
i @ his
a180
( j a 200se t
SQL> set linesize
SQL> set pagesize
b
SQL> b ia to u
i o A plan_table_output from
SQL> select
n
ntotable(dbms_xplan.display('plan_table',null,'ALL'));
e A
J o s PLAN_TABLE_OUTPUT
--------------------------------------------------------------------
--------------------------------------------------------------------
--------------------------------------------
--------------------------------------------------------------------
--------------------------------------
--------------------------------------------------------------------
--------------------------------------
1 | 1764 | 7316P (1)|999:59:59 |
| 1 | RESULT CACHE | b7rh5vw33py4ug4n8xaav6525g |
| | | |
1 | 1764 | | |
| 3 | MERGE JOIN CARTESIAN | |
294P| 15E| 7316P (1)|999:59:59 |

941T| 1264P| 23P (1)|999:59:59 |
3007G| 3446T| 74T (1)|999:59:59 |
9606M| 9018G| 238G (1)|999:59:59 |

30M| 21G| 761M (1)|999:59:59 |
| 8 | MERGE JOIN CARTESIAN| |
98011 | 47M| 2432K (1)| 08:06:26 |
313 | 78876 | 7748 (1)| 00:01:33 |
313 | 78876 | 2424K (1)| 08:04:53 |
313 | 78876 | 7746 (1)| 00:01:33 |
n s e
i ce
313 | 78876 | 761M (1)|999:59:59 |
b l el
|* 13 | TABLE ACCESS FULL | CACHEJFV
313 | 78876 | 7746 (1)| 00:01:33 |
|
fer a
| 14 | BUFFER SORT |
a n s
|
313 | 78876 | 238G (1)|999:59:59 |
n -tr
o
an
313 | 78876 | 7746 (1)| 00:01:33 |
| 16 | BUFFER SORT |
h s
a deฺ |
313 | 78876 | 74T (1)|999:59:59 |
e )
s Gui
r ฺ
mi dent
313 | 78876 | 7746 ฺ
(1)| 00:01:33 |
p
g Stu
| 18 | BUFFER SORT
d | |
313 | 78876 |
|* 19 | b a@ this
23P (1)|999:59:59 |
i
TABLE ACCESS FULL | CACHEJFV |
313 | 78876 | 7746
i a (ja use
(1)| 00:01:33 |
| 20 |
A b to
BUFFER SORT | |
o n io
313 | 78876 | 7316P (1)|999:59:59 |
nt
e A 313 | 78876 | 7746 (1)| 00:01:33 |
J o s --------------------------------------------------------------------
--------------------------------------

-------------------------------------------------------------
1 - SEL$1
9 - SEL$1 / C7@SEL$1
11 - SEL$1 / C6@SEL$1
13 - SEL$1 / C5@SEL$1
15 - SEL$1 / C4@SEL$1
17 - SEL$1 / C3@SEL$1
19 - SEL$1 / C2@SEL$1
21 - SEL$1 / C1@SEL$1

---------------------------------------------------
9 - filter("C7"."C"='b')
11 - filter("C6"."C"='b')
13 - filter("C5"."C"='b')

15 - filter("C4"."C"='b')
17 - filter("C3"."C"='b')
19 - filter("C2"."C"='b')
21 - filter("C1"."C"='b')
Column Projection Information (identified by operation id):

-----------------------------------------------------------
1 - COUNT(*)[22]
2 - (#keys=0) COUNT(*)[22]
"C5"."C"[VARCHAR2,500],
"C5"."C"[VARCHAR2,500],
n s e
i ce
b l el
"C5"."C"[VARCHAR2,500],
fer a
a n s
-tr
n
o
an
"C5"."C"[VARCHAR2,500],
h s
a deฺ
e )
s Gui
r ฺ
mi dent
8 - (#keys=0) "C7"."C"[VARCHAR2,500], "C6"."C"[VARCHAR2,500]
9 - "C7"."C"[VARCHAR2,500]
p ฺ
g Stu
d
10 - (#keys=0) "C6"."C"[VARCHAR2,500]
b i a@ this
11 - "C6"."C"[VARCHAR2,500]
12 - (#keys=0) "C5"."C"[VARCHAR2,500]
i a (ja use
13 - "C5"."C"[VARCHAR2,500]
b to
14 - (#keys=0) "C4"."C"[VARCHAR2,500]
A
o n io
15 - "C4"."C"[VARCHAR2,500]
16 - (#keys=0) "C3"."C"[VARCHAR2,500]
nt
e A 17 - "C3"."C"[VARCHAR2,500]
J o s 18 - (#keys=0) "C2"."C"[VARCHAR2,500]
19 - "C2"."C"[VARCHAR2,500]
20 - (#keys=0) "C1"."C"[VARCHAR2,500]
21 - "C1"."C"[VARCHAR2,500]
Result Cache Information (identified by operation id):

------------------------------------------------------
1 - column-count=1; dependencies=(QRC.CACHEJFV);
attributes=(single-row); parameters=(nls); name="select count(*)
where c1.c='b' and c2."
Note
-----
89 rows selected.

Elapsed: 00:00:00.07
SQL>
COUNT(*)
----------
1
Elapsed: 00:00:00.04
SQL>
n s e
i ce
TYPE STATUS NAME
b l el
OBJECT_NO ROW_COUNT
---------- --------- -----------------------------------------------
fer a
n s
--------------------------------------------------------------------
a
------------- ---------- ----------
o n -tr
an
ROW_SIZE_AVG
------------
h s
a deฺ
71474 0
e )
s Gui
0 r ฺ
i ent
p ฺ m d
Result Invalid
d g
select
t u
/*+ result_cache
S
q_name(Q1) */ count(*)
0 1
b i a@fromthcachejfv
is c1,cachejfv c2,cachejfv
c3,cachejfv ( j a
c4,cachejfv s e c5,cachejfv c6, cac
b i a5
to u
i o A
n
toResult Published select /*+ result_cache q_name(Q1) */ count(*)
A n 0 1
o s e from cachejfv c1,cachejfv c2,cachejfv

J c3,cachejfv c4,cachejfv c5,cachejfv c6, cac
5
Elapsed: 00:00:00.00
SQL>
SQL> alter session set result_cache_mode=manual;
Session altered.
Elapsed: 00:00:00.00
SQL>
SQL>
--------------------------------------------------------------
set echo on
show parameter result_cache_mode

alter session set result_cache_mode=force;
explain plan for

select count(*)
set linesize 180

set pagesize 200
select plan_table_output from

n s e
i ce
b l el
select count(*)
fe r a
a n s
o n -tr
an
s
a deฺ
v$result_cache_objects order by 1; ) h
r ฺ e s Gui
ฺ mi dent
alter session set result_cache_mode=manual;
p
d g Stu
i @ V$RESULT_CACHE_OBJECTS
aQuery h is
a b
16) Clear the result cache.
e t to verify the clear
operation. a (j u s
b i t o
i o A dbms_result_cache.flush;
SQL> exec
n
to
e An PL/SQL procedure successfully completed.
Jos Elapsed: 00:00:00.01
no rows selected
Elapsed: 00:00:00.00
SQL>
17) Create a PL/SQL function that uses the result cache by running the cre_func.sql
script.
SQL> @cre_func
SQL> create or replace function CACHEJFV_COUNT(v varchar2)
2 return number
3 result_cache relies_on (cachejfv)
4 is
5 cnt number;
6 begin
7 select count(*) into cnt

9 where c1.c=v and c2.c=v and c3.c=v and c4.c=v and c5.c=v and
c6.c=v and c7.c=v;
10 return cnt;
11 end;
12 /
Function created.
Elapsed: 00:00:00.15
SQL>
--------------------------------------------------------------
create or replace function CACHEJFV_COUNT(v varchar2)

n s e
return number
i ce
result_cache relies_on (cachejfv)
b l el
is
cnt number;
fe r a
begin
a n s
select count(*) into cnt
o n -tr
an
h s
a deฺ
where c1.c=v and c2.c=v and c3.c=v and c4.c=v and c5.c=v and c6.c=v
and c7.c=v;
e )
s Gui
r ฺ
mi dent
return cnt;
end;
p ฺ
g Stu
/
d
b i a@ this
ain the result
18) Determine what(jis se cache by querying V$RESULT_CACHE_OBJECTS.
i a to u
b type,status,name,object_no,row_count,row_size_avg
i o A
SQL> select from
n
Anto
e no rows selected
Jos Elapsed: 00:00:00.00
SQL>
19) Call the new function with ‘b’ as its argument. What do you observe?
a) It takes a long time to execute because the result is not cached yet. After
executing the function, the function’s result for the ‘b’ argument is cached.
SQL> select cachejfv_count('b') from dual;
CACHEJFV_COUNT('B')
-------------------
1
Elapsed: 00:00:01.56
SQL>

TYPE STATUS NAME
OBJECT_NO ROW_COUNT
---------- --------- -----------------------------------------------
--------------------------------------------------------------------
------------- ---------- ----------
ROW_SIZE_AVG
------------
Dependency Published QRC.CACHEJFV_COUNT
71475 0
0

71474 0
0
Result Published
n s e
"QRC"."CACHEJFV_COUNT"::8."CACHEJFV_COUNT"#8440831613f0f5d3 #1
i ce
0 1
b l el
4
fe ra
a n s
Elapsed: 00:00:00.00
o n -tr
an
SQL>
SQL>
h s
a What edoฺ you observe?
20) Call the new function with ‘b’ as its argument )
s Guagain. i d
a) This time the function executes m i
almostrฺeinstantaneously.
nt
p ฺ
g Sfrom d e
d
SQL> select cachejfv_count('b') tu dual;
b i a@ this
i a (ja use
CACHEJFV_COUNT('B')
-------------------
A b to1
ni o
t o
An SQL>
Elapsed: 00:00:00.00
e
Jos 21) Call the new function with ‘c’ as its argument again. What do you observe?
a) Again it takes a long time to execute the function because of the new value for the
argument. After execution, the second result is cached.
SQL> select cachejfv_count('c') from dual;
CACHEJFV_COUNT('C')
-------------------
1
Elapsed: 00:00:04.15
TYPE STATUS NAME

OBJECT_NO ROW_COUNT
---------- --------- -----------------------------------------------
--------------------------------------------------------------------
------------- ---------- ----------

ROW_SIZE_AVG
------------
Dependency Published QRC.CACHEJFV_COUNT
71475 0
0

71474 0
0
Result Published
0 1
4
Result Published
n s e
i ce
0 1
b l el
4
fer a
a n s
Elapsed: 00:00:00.01
n -tr
o
an
SQL>
SQL>
h
SQL> select cachejfv_count('c') from dual; s
a deฺ
e )
s Gui
CACHEJFV_COUNT('C') r ฺ
i ent
-------------------
p ฺ m
1
d g Stud
b i
SQL>
i a (ja use
A b to
i o
nton
e A
J o s

n s e
i ce
b l el
fer a
a n s
n - tr
o
s an
) h a deฺ
r ฺ e s Gui
p ฺ mi dent
d g Stu
b i a@ this
i a (ja use
Ab to
o n io
nt
e A
J o s

Practice 5-1: Extracting Execution Plans
In this practice, you use various methods to extract the execution plan used by the
optimizer to execute a query. Note that all scripts needed for this lab can be found in your
$HOME/solutions/Explain_Plan directory.
1) Connected as the oracle user from a terminal session, execute the ep_setup.sh
script. This script creates a new user called EP and a table called TEST used
throughout this lab.
[oracle@edrsr33p1-orcl Explain_Plan]$ ./ep_setup.sh

n s e
ice
Connected to:
b l el
fe ra
Production
a n s
-tr
o n
SQL>
s an
SQL> drop user ep cascade;
) h a deฺ
drop user ep cascade
r ฺ e s Gui
*
ERROR at line 1: p ฺ mi dent
d
ORA-01918: user 'EP' does g not S tu
exist
b i a@ this
SQL>
i a (ja use
A b usertoep identified by ep default tablespace users
SQL> create
n i o
nt o
e A User created.
J o s SQL>
SQL> grant connect, resource, dba to ep;
Grant succeeded.
SQL>
SQL> connect ep/ep
Connected.
SQL>
SQL> drop table test purge;
drop table test purge
*
ERROR at line 1:
SQL>
SQL> create table test(c number, d varchar2(500));
Table created.

Practice 5-1: Extracting Execution Plans (continued)
SQL>
SQL> begin
3 insert into test
values(1,'aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa');
4 end loop;
5 commit;
6 end;
7 /
SQL>
n s e
SQL> create index test_c_indx on test(c);
i ce
b l el
Index created.
fer a
SQL>
a n s
SQL> exec dbms_stats.gather_schema_stats('EP');
n -tr
o
s an
) h a deฺ
SQL>
r ฺ e s Gui
ฺ m i ent
System altered.
d p
g Stud
SQL> b i a@ this
i a (ja flush
SQL> alter system
u se buffer_cache;
A b
altered. t
o
ni o
System
ntoSQL>
e A
J o s SQL> set echo off
[oracle@edrsr33p1-orcl Explain_Plan]$
--------------------------------------------------------------
#!/bin/bash
cd /home/oracle/solutions/Explain_Plan
export

sqlplus / as sysdba @ep_setup.sql
--------------------------------------------------------------
set echo on
drop user ep cascade;
create user ep identified by ep default tablespace users temporary

tablespace temp;
grant connect, resource, dba to ep;
connect ep/ep
drop table test purge;

n s e
i ce
create table test(c number, d varchar2(500));
b l el
begin
fe r a
a n s
insert into test
n - tr
o
an
values(1,'aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa');
h s
a deฺ
end loop;
e )
s Gui
r ฺ
mi dent
commit;
end;
p ฺ
g Stu
/
d
b i a@ ton
create index test_c_indx h istest(c);
i a (ja use
b to
exec dbms_stats.gather_schema_stats('EP');
A
o
ni system flush shared_pool;
n t oalter
e A
Jos
set echo off

set term off
select count(*) from test t1, test t2 where t1.c=t2.c and t1.c=1;
exit;
2) From the same terminal session (referred to as session 1 in the rest of this lab), be
ready to execute the ep_session_issue.sh script. Enter the command, but do
not execute it yet.
Session 1:
----------
[oracle@edrsr33p1-orcl Explain_Plan]$ ./ep_session_issue.sh
--------------------------------------------------------------

#!/bin/bash
export
sqlplus ep/ep @ep_session_issue.sql
--------------------------------------------------------------
n s e
set echo off
i ce
set termout off
b l el
alter session set optimizer_mode=rule;
fe r a
a n s
set termout on
o n -tr
an
set echo on
set timing on
h s
a deฺ
e )
st2 where i
ut1.c=t2.c
select count(*) from test t1, test
i r ฺ n t G and t1.c=1;
m
exit;
d gpฺ Stude
b i a@ this
ja usession
3) From a second (terminal
a se (referred to as session 2 in the rest of this lab),
connect as i to user. After this, connect to a SQL*Plus session as the SYS
b the oracle
i
user.o A
From that SQL*Plus session, be ready to use SQL Monitoring to monitor the
t o n
e An execution plan used by session 1. You can execute the ep_monitoring.sql
script for that purpose. Enter the command, but do not execute it yet. Note: Ensure
Jos that you understand the coordination between both sessions by pre-reading steps 4
and 5 before you continue.
Session 2:
----------
[oracle@edrsr33p1-orcl Explain_Plan]$ sqlplus / as sysdba
Connected to:
Production
SQL> @ep_monitor

--------------------------------------------------------------
set echo on
set long 10000000
set longchunksize 10000000
set linesize 200

set pagesize 1000
exec dbms_lock.sleep(8);
select
dbms_sqltune.report_sql_monitor(sql_id=>'dkz7v96ym42c6',report_level
=>'ALL') from dual;
n s e
4) After you are ready in both the sessions, press [Enter] in session 1 to start the
ice
execution of the ep_session_issue.sh script. Note: Do not wait. Proceed with
b l el
the next step immediately.
fe ra
a n s
-tr
Session 1:
----------
o n
s an
[oracle@edrsr33p1-orcl Explain_Plan]$ ./ep_session_issue.sh
SQL*Plus: Release 11.1.0.6.0 - Production
h a on Wedd ฺ 2 20:12:47 2008
eApr
) i
Copyright (c) 1982, 2007, Oracle.i r ฺesAllnrights
t Gu reserved.
g p ฺm ude
d S t
@ s
thi Edition Release 11.1.0.6.0 -
ia Enterprise
Connected to:
Oracle Database b
Production ( j a 11g
e
b a to us Oracle Label Security, OLAP, Data Mining
iPartitioning,
A Application Testing options
With the
andoReal
i
to n
A n SQL>
o s e SQL> set timing on

J SQL>
SQL> select count(*) from test t1, test t2 where t1.c=t2.c and
t1.c=1;
5) In session 2, enter return to start the execution of the ep_monitor.sql script.

After the execution, enter “/” and go back to your SQL*Plus session as many times
as necessary until session 1 is done with its execution. What do you observe?
a) You can see that session 1 uses NESTED LOOPS on top of two INDEX RANGE
SCANS to execute the query. It takes approximately 47 seconds to execute session
1’s query. The time depends on your environment. The big advantage of SQL
Monitoring is that you can clearly see which steps in the execution plan take most
of the resources. In this case, you clearly see that you do only one scan of the
index, and that for each row returned, you execute another index scan to probe.
This is not really efficient. Also there is no costs information for this monitored
plan.

Session 2:
----------
SQL> @ep_monitor
SQL> set long 10000000
SQL> set longchunksize 10000000

SQL>
SQL> exec dbms_lock.sleep(8);
SQL>
SQL> select
n s e
=>'ALL') from dual;
i ce
b l el
DBMS_SQLTUNE.REPORT_SQL_MONITOR(SQL_ID=>'DKZ7V96YM42C6',SESSION_ID=>
f e ra
ns
:SESSID,REPORT_LEVEL=>'ALL')
- t r a
--------------------------------------------------------------------
n
--------------------------------------------------------------------
o
an
----------------------------------------------------------------
SQL Monitoring Report
h s
a deฺ
SQL Text
e )
s Gui
r ฺ
mi dent
--------------------------------------------------------------------
p ฺ
--------------------------------------------------------------------
dg S tu t2 where t1.c=t2.c and t1.c=1
-------------------------------------------
select count(*) from@
b i a t h is
test t1, test
(ja use
--------------------------------------------------------------------
b a
--------------------------------------------------------------------
i to
-------------------------------------------
o A
t o ni Information
Global
An Status : EXECUTING
e
Jos
Instance ID : 1
Session ID : 138
SQL ID : dkz7v96ym42c6
SQL Execution ID : 16777222
Plan Hash Value : 1643938535
Execution Started : 04/02/2008 20:12:46
First Refresh Time : 04/02/2008 20:12:54
Last Refresh Time : 04/02/2008 20:12:56
-----------------------------------------
| Elapsed | Cpu | Other | Buffer |
| Time(s) | Time(s) | Waits(s) | Gets |
-----------------------------------------
| 8.10 | 8.09 | 0.01 | 134K |
-----------------------------------------
SQL Plan Monitoring Details

====================================================================
=================================================================

| Id | Operation | Name | Rows | Cost |
Time | Start | Starts | Rows | Activity | Activity Detail |
| | | | (Estim) | |
Active(s) | Active | | (Actual) | (percent) | (sample #)
|
====================================================================
=================================================================
| 0 | SELECT STATEMENT | | | |
| | 1 | | | |
| 1 | SORT AGGREGATE | | | |
1 | +6 | 1 | 0 | 12.50 | Cpu (1) |
| -> 2 | NESTED LOOPS | | | |
3 | +8 | 1 | 79096K | | |
| -> 3 | INDEX RANGE SCAN | TEST_C_INDX | | |
3 | +8 | 1 | 3954 | | |
n s e
9 | +2 | 3955 | 79096K | 87.50 | Cpu (7) |
i ce
====================================================================
b l el
=================================================================
fer a
a n s
SQL>
o n -tr
an
SQL> /
h s
a deฺ
e )
s Gui
r ฺ
i ent
--------------------------------------------------------------------
ฺ m
--------------------------------------------------------------------
p d
d g t u
----------------------------------------------------------------
S
b i a@ this
SQL Text
i a (ja use
b to
--------------------------------------------------------------------
A
i o
--------------------------------------------------------------------
n
nto-------------------------------------------
e A select count(*) from test t1, test t2 where t1.c=t2.c and t1.c=1
J o s --------------------------------------------------------------------
--------------------------------------------------------------------
-------------------------------------------
Global Information
Status : EXECUTING
Instance ID : 1
Session ID : 138
-----------------------------------------
-----------------------------------------
| 10 | 10 | 0.01 | 167K |
-----------------------------------------


====================================================================
=================================================================

| | | | (Estim) | |
|
====================================================================
=================================================================
| | 1 | | | |
1 | +6 | 1 | 0 | 10.00 | Cpu (1) |
n s e
| -> 2 | NESTED LOOPS | | | |
i ce
5 | +8 | 1 | 96151K | | |
b l el
| -> 3 |
5 | +8 |
INDEX RANGE SCAN | TEST_C_INDX |
1 | 4807 | |
| |
fe
| r a
| -> 4 | INDEX RANGE SCAN | TEST_C_INDX | |
a n s |
11 | +2 | 4808 | 96151K | 90.00 | Cpu (9)
o n -tr |
an
====================================================================
=================================================================
h s
a deฺ
e )
s Gui
SQL> / r ฺ
i ent
p ฺ m d
d g t u
S
b i a@ this
--------------------------------------------------------------------
i a (ja use
--------------------------------------------------------------------
b to
----------------------------------------------------------------
A
o
ni Text
ntoSQL
e A
J o s --------------------------------------------------------------------
--------------------------------------------------------------------
-------------------------------------------
select count(*) from test t1, test t2 where t1.c=t2.c and t1.c=1
--------------------------------------------------------------------
--------------------------------------------------------------------
-------------------------------------------
Global Information
Status : EXECUTING
Instance ID : 1
Session ID : 138
-----------------------------------------

-----------------------------------------
| 12 | 12 | 0.01 | 201K |
-----------------------------------------

====================================================================
=================================================================
| | | | (Estim) | |
|
====================================================================
=================================================================
n s e
i ce
| | 1 | | | |
b l el
|
1 |
1 | SORT AGGREGATE
+6 | 1 |
|
0 |
|
8.33 | Cpu (1)
| |
fe
| r a
| -> 2 | NESTED LOOPS | | |
a n s |
7 | +8 | 1 | 113M | |
o n -tr |
an
7 | +8 | 1 | 5664 | | |
| -> 4 |
h s
a deฺ
INDEX RANGE SCAN | TEST_C_INDX | | |
13 | +2 | 5665 | 113M |
e )
s Gui
91.67 | Cpu (11) |
r ฺ
mi dent
====================================================================
p ฺ
=================================================================
g Stu
d
SQL> / b i a@ this
i a (ja use
b to
A
i o
n
nto--------------------------------------------------------------------
e A --------------------------------------------------------------------
J o s ----------------------------------------------------------------
SQL Text
--------------------------------------------------------------------
--------------------------------------------------------------------
-------------------------------------------
--------------------------------------------------------------------
--------------------------------------------------------------------
-------------------------------------------
Global Information
Status : EXECUTING
Instance ID : 1
Session ID : 138

-----------------------------------------
-----------------------------------------
| 20 | 20 | 0.04 | 334K |
-----------------------------------------

====================================================================
=================================================================
| | | | (Estim) | |
n s e
i ce
|
b l el
====================================================================
fe
================================================================= r a
a n|s
| | 1 | | |
n - tr |
o
an
1 | +6 | 1 | 0 | 5.00 | Cpu (1) |
| -> 2 | NESTED LOOPS |
h |s
a deฺ | |
15 | +8 | 1 | 182M |
e )
s Gui| |
r ฺ
mi dent
15 | +8 | 1 | ฺ
9082 |
p
g Stu
| |
| -> 4 |
d
21 | +2 | 9083 |
b i a@ this
182M | 95.00 | Cpu (19) |
====================================================================
i a (ja use
=================================================================
Ab to
o n io
SQL> /
nt
e A
J o s DBMS_SQLTUNE.REPORT_SQL_MONITOR(SQL_ID=>'DKZ7V96YM42C6',SESSION_ID=>
--------------------------------------------------------------------
--------------------------------------------------------------------
----------------------------------------------------------------
SQL Text
--------------------------------------------------------------------
--------------------------------------------------------------------
-------------------------------------------
--------------------------------------------------------------------
--------------------------------------------------------------------
-------------------------------------------
Global Information
Status : EXECUTING
Instance ID : 1
Session ID : 138

-----------------------------------------
-----------------------------------------
| 28 | 28 | 0.05 | 468K |
-----------------------------------------

====================================================================
n s e
=================================================================
i ce
b l el
Time
| |
| Start | Starts | Rows
|
| Activity | Activity Detail |
| (Estim) | |
fer a
Active(s) | Active | | (Actual) | (percent) | n s
(sample #)
a
|
n - tr
o
an
====================================================================
=================================================================
h |s
a deฺ | |
| | 1 | |
e )
s Gui
| |
r ฺ
mi dent
1 | +6 | 1 |
p ฺ
0 |
g Stu
3.45 | Cpu (1) |
| -> 2 | NESTED LOOPS
d | | | |
24 |
| -> 3 |
+8 |
b
1 |
i a@ this
250M | |
|
24 | +8 |
i a (ja use
1 | 12518 | | |
| -> 4 |
Ab to
o n io
30 | +2 | 12519 | 250M | 96.55 | Cpu (28) |
====================================================================
nt
e A =================================================================
J o s
SQL> /
--------------------------------------------------------------------
--------------------------------------------------------------------
----------------------------------------------------------------
SQL Text
--------------------------------------------------------------------
--------------------------------------------------------------------
-------------------------------------------
--------------------------------------------------------------------
--------------------------------------------------------------------
-------------------------------------------
Global Information

Status : EXECUTING
Instance ID : 1
Session ID : 138

-----------------------------------------
-----------------------------------------
| 40 | 40 | 0.06 | 669K |
-----------------------------------------
n s e
i ce
b l el
fer
==================================================================== a
n s
=================================================================
a
| Id | Operation | Name | Rows
n - tr
| Cost |
o
an
| | | | (Estim) | |
Active(s) | Active |
h s
a deฺ
| (Actual) | (percent) | (sample #)
|
e )
s Gui
r ฺ
mi dent
====================================================================
p ฺ
=================================================================
g Stu
|
d
0 | SELECT STATEMENT | | | |
|
| 1 |
| 1 |
b i a@ this
SORT AGGREGATE
|
|
|
| |
|
|
35 | +6 |
i a (ja use
1 | 0 | 4.88 | Cpu (2) |
| -> 2 |
Ab to
NESTED LOOPS | | | |
o n io
36 |
| -> 3 |
+8 | 1 | 353M | |
|
nt
e A 36 | +8 | 1 | 17663 | | |
J o s | -> 4 |
42 |
INDEX RANGE SCAN | TEST_C_INDX |
+2 | 17664 | 353M | 95.12 | Cpu (39)
| |
|
====================================================================
=================================================================
SQL> /
--------------------------------------------------------------------
--------------------------------------------------------------------
----------------------------------------------------------------
SQL Text
--------------------------------------------------------------------
--------------------------------------------------------------------
-------------------------------------------

--------------------------------------------------------------------
--------------------------------------------------------------------
-------------------------------------------
Global Information
Status : DONE (ALL ROWS)
Instance ID : 1
Session ID : 138
-------------------------------------------------
n s e
| Elapsed | Cpu | Other | Fetch | Buffer |
ice
| Time(s) | Time(s) | Waits(s) | Calls | Gets |
b l el
-------------------------------------------------
| 46 | 46 | 0.06 | 1 | 760K |
fe r a
-------------------------------------------------
a n s
o n -tr
s an
h a deฺ
====================================================================
)
s G| uRows i
===============================================================
e
| rฺName
| iActivity t
| Id | Operation | Cost | Time
| Start | Starts | Rows ฺm n
p | tude | (Estim) |
| Activity Detail |
| |
Active(s) | Active |@ d g | S
|
| b i a t h is (Actual) | (percent) | (sample #)
i a (ja use
====================================================================
b to
===============================================================
A
1ni|
o
nto| 1 | SORT AGGREGATE |

+48 | 1 | 1 | | |
e A | | |
J o s 43 |
| 2 |
+6 | 1 |
NESTED LOOPS |
1 | 4.35 | Cpu (2)
| | |
|
41 | +8 | 1 | 400M | | |
| 3 | INDEX RANGE SCAN | TEST_C_INDX | | |
41 | +8 | 1 | 20000 | | |
| 4 | INDEX RANGE SCAN | TEST_C_INDX | | |
47 | +2 | 20000 | 400M | 95.65 | Cpu (44) |
====================================================================
===============================================================
SQL>
6) After approximately 47 seconds (depending on your environment), you should see the
following output in your session 1:
Session 1:
----------

…
COUNT(*)
----------
400000000
Elapsed: 00:00:47.21
SQL>
SQL> exit;
n s e
i ce
7) From session 1, connect as the EP user in the SQL*Plus session.
b l el
Session 1:
fe ra
----------
a n s
o
[oracle@edrsr33p1-orcl Explain_Plan]$ sqlplus ep/ep n -tr
n
SQL*Plus: Release 11.1.0.6.0 - Production s onaWed Apr 2 20:14:03 2008
) h a deฺ
Copyright (c) 1982, 2007, Oracle.ฺes
r G uireserved.
All rights
p ฺ mi dent
d g Stu
a@
Connected to:
b i
Oracle Database 11g
t h is Edition Release 11.1.0.6.0 -
Enterprise
Production
a (ja useOracle Label Security, OLAP, Data Mining
With the iPartitioning,
A
and Real to Testing options
b Application
ni o
ntoSQL>
e A
J o s 8) Use PLAN_TABLE to determine the execution plan of the query that was executed in
step 4. What do you observe?
a) This time the execution plan uses a hash join on top of two index fast full scans.
Session 1:
----------
SQL> @ep_explain
SQL>
SQL>
2 select count(*) from test t1, test t2 where t1.c=t2.c and
t1.c=1;
Explained.
SQL>
SQL> select * from table(dbms_xplan.display);

PLAN_TABLE_OUTPUT
--------------------------------------------------------------------
--------------------------------------------------------------------
----------------------------------------------------------------
--------------------------------------------------------------------
------------------
(%CPU)| Time |
--------------------------------------------------------------------
------------------
(99)| 00:00:22 |
| 1 | SORT AGGREGATE | | 1 | 6 |
n s e
| |
i ce
|* 2 | HASH JOIN | | 400M| 2288M| 2131
b l el
(99)| 00:00:22 |
|* 3 | INDEX FAST FULL SCAN| TEST_C_INDX | 20000 | 60000 | 13
fe r a
(0)| 00:00:01 |
a n s
|* 4 | -
INDEX FAST FULL SCAN| TEST_C_INDX | 20000 | 60000 |
n tr
13
o
an
(0)| 00:00:01 |
--------------------------------------------------------------------
------------------
h s
a deฺ
e )
s Gui id):
r ฺ
mi dent
p ฺ
---------------------------------------------------
g Stu
d
b i a@ this
2 - access("T1"."C"="T2"."C")
3 - filter("T1"."C"=1)
i a (ja use
4 - filter("T2"."C"=1)
A bselected.to
ni o
18 rows
ntoSQL>
e A
J o s --------------------------------------------------------------
set echo on
explain plan for

select * from table(dbms_xplan.display);
9) Now, you want to monitor the previous execution plan to compare it with the one
generated in step 4. In addition, you want to make sure you use the correct plan this
time. So, in your session 1, start autotrace, and be ready to execute the following
query. Do not execute it yet as you need to start SQL Monitoring in your session 2:

Session 1:
----------

SQL> @ep_execute
10) From your session 2, be ready to execute your SQL Monitoring command again. Do
not execute it yet though.
Session 2:
----------
SQL> @ep_monitor
n s e
i ce
--------------------------------------------------------------
b l el
fe r a
set echo on
a n s
-tr
set long 10000000
o n
an
set linesize 200
set pagesize 1000
h s
a deฺ
e )
s Gui
r ฺ
select
p ฺ mi dent
dg Stu
=>'ALL') from dual; @
b i a t h is
i a (ja use
execution tofoyour query from session 1 by pressing [Enter]. Note: Move to
11) Start the b
i
next o A
stepwithout waiting.
t o n
e An
Jos
Session 1:
----------
SQL> @ep_execute
SQL> set echo on
SQL>
SQL> set timing on
SQL>
SQL> select count(*) from test t1, test t2 where t1.c=t2.c and
t1.c=1;
--------------------------------------------------------------
set echo on
set timing on

12) From your session 2, start monitoring your query by pressing the return key. After the
query is executed, enter “/” and go back to your SQL*Plus session as many times as
necessary until session 1 is done with its execution. What do you observe?
a) You can see that the optimizer uses a hash join on top of two index fast full scans.
Looking at the various reports, you can clearly see how the optimizer processes a
hash join by reading the driving index in memory first. This operation is quick.
Though you cannot see it run, it is already done the first time you can look at it.
Then the probe in performed on the index again. This operation takes more time.
Also note that cost information is provided in the execution plan.
Session 2:
----------
n s e
ice
el
SQL> @ep_monitor
SQL> set echo on
a b l
SQL> set long 10000000
fe r
SQL> set longchunksize 10000000
a n s
-tr
o n
an
SQL>
SQL>exec dbms_lock.sleep(8);
h s
a deฺ
e
PL/SQL procedure successfully completed.)
s Gui
r ฺ
SQL> p ฺ mi dent
SQL> select d g Stu
=>'ALL') from a b ia@ this
(j se
dual;
b i a to u
A
o
ni
to--------------------------------------------------------------------
A n --------------------------------------------------------------------
o s e ----------------------------------------------------------------
J SQL Monitoring Report
SQL Text
--------------------------------------------------------------------
--------------------------------------------------------------------
-------------------------------------------
--------------------------------------------------------------------
--------------------------------------------------------------------
-------------------------------------------
Global Information
Status : EXECUTING
Instance ID : 1
Session ID : 124

-----------------------------------------
-----------------------------------------
| 6.29 | 6.29 | 0.00 | 8 |

-----------------------------------------

====================================================================
====================================================================
==========
Time | Start | Starts | Rows | Memory | Activity | Activity
n s e
Detail |
i ce
| | | | (Estim) | |
b l el
Active(s) | Active |
(sample #) |
| (Actual) | | (percent) |
fer a
n s
====================================================================
a
- tr
====================================================================
n
o
an
==========
| 0 | SELECT STATEMENT | | | 2131 |
| | 1 | | |
h s
a deฺ |
|
e )
s Gui
r ฺ
mi dent
| 1 | SORT AGGREGATE | | 1 | |
8 | +0 | 1 |
p ฺ
0 |
g Stu
| 25.00 | Cpu (2)
|
d
| -> 2 |
8 | +1 |
HASH JOIN
b i
1 | a@ this |
113M | 1138K |
| 400M | 2131 |
75.00 | Cpu (6)
|
i a (ja use
| -> 3 |
Ab to
INDEX FAST FULL SCAN | TEST_C_INDX | 20000 | 13 |
o
1 |
|n io +8 | 1 | 20000 | | |
nt
e A | -> 4 | INDEX FAST FULL SCAN | TEST_C_INDX | 20000 | 13 |
J o s 1 |
|
+8 | 1 | 5632 | | |
====================================================================
====================================================================
==========
SQL>
SQL> /
--------------------------------------------------------------------
--------------------------------------------------------------------
----------------------------------------------------------------
SQL Text
--------------------------------------------------------------------
--------------------------------------------------------------------
-------------------------------------------

--------------------------------------------------------------------
--------------------------------------------------------------------
-------------------------------------------
Global Information
Status : EXECUTING
Instance ID : 1
Session ID : 124
n s e
-----------------------------------------
i ce
b l el
-----------------------------------------
fer a
| 7.86 | 7.86 | 0.00 | 10 |
a n s
-----------------------------------------
o n -tr
s an
) h a deฺ
r ฺ e s Gui
====================================================================
========== ฺ m i ent
====================================================================
| Id |
d g
Operation p t u|d Name | Rows | Cost |
Time
a@ this
| Start | Starts | S
Rows | Memory | Activity | Activity
Detail | b i
| |
a ja use
(Active | | (Estim) | |
Active(s) i
b#) | to
| | | (Actual) | | (percent) |
i o A
(sample
n
to====================================================================
A n ====================================================================
o s e ==========
J |
|
| 1 | |
|
|
|
|
| 2131 |
|
8 | +0 | 1 | 0 | | 20.00 | Cpu (2)
|
| -> 2 | HASH JOIN | | 400M | 2131 |
10 | +1 | 1 | 133M | 1138K | 80.00 | Cpu (8)
|
| 3 | INDEX FAST FULL SCAN | TEST_C_INDX | 20000 | 13 |
1 | +8 | 1 | 20000 | | |
|
| -> 4 | INDEX FAST FULL SCAN | TEST_C_INDX | 20000 | 13 |
3 | +8 | 1 | 6656 | | |
|
====================================================================
====================================================================
==========

SQL> /
--------------------------------------------------------------------
--------------------------------------------------------------------
----------------------------------------------------------------
SQL Text
--------------------------------------------------------------------
--------------------------------------------------------------------
-------------------------------------------
--------------------------------------------------------------------
--------------------------------------------------------------------
n s e
-------------------------------------------
i ce
b l el
Global Information
Status : EXECUTING
fer a
Instance ID : 1
a n s
Session ID : 124
n -tr
o
an
h s
a deฺ
Execution Started : )
s Gui
04/02/2008 20:14:41
e
r ฺ
mi dent
Last Refresh Time : ฺ
04/02/2008 20:14:55
p
g Stu
d
| Elapsed | b
Cpu a@ t
| Other is | Buffer |
-----------------------------------------
i h
| Time(s) | (
a ja u|seWaits(s) | Gets |
Time(s)
b i to12 | 0.00 | 15 |
-----------------------------------------
o A
ni
| 12 |
nto-----------------------------------------
e A
J o s SQL Plan Monitoring Details
====================================================================
====================================================================
==========
Detail |
| | | | (Estim) | |
Active(s) | Active | | (Actual) | | (percent) |
(sample #) |
====================================================================
====================================================================
==========
| | 1 | | | |
|
8 | +0 | 1 | 0 | | 14.29 | Cpu (2)
|

| -> 2 | HASH JOIN | | 400M | 2131 |
14 | +1 | 1 | 184M | 1138K | 85.71 | Cpu (12)
|
1 | +8 | 1 | 20000 | | |
|
| -> 4 | INDEX FAST FULL SCAN | TEST_C_INDX | 20000 | 13 |

7 | +8 | 1 | 9216 | | |
|
====================================================================
====================================================================
==========
SQL> /
n s e
i ce
b l el
fer a
--------------------------------------------------------------------
--------------------------------------------------------------------
n s
----------------------------------------------------------------
a
o n -tr
SQL Text
s an
h a deฺ
--------------------------------------------------------------------
)
r ฺ e s Gui
--------------------------------------------------------------------
select count(*) from test t1,ฺ m i t2enwhere

-------------------------------------------
test t t1.c=t2.c and t1.c=1
d p
g Stu d
--------------------------------------------------------------------
b i a@ this
--------------------------------------------------------------------
-------------------------------------------
i a (ja use
b to : EXECUTING
Global Information
A
i o
Status
ntonInstance ID : 1
e A Session ID : 124
J o s SQL ID
SQL Execution ID
:
:
dkz7v96ym42c6
16777223
-----------------------------------------
-----------------------------------------
| 22 | 22 | 0.01 | 28 |
-----------------------------------------

====================================================================
====================================================================
==========

Detail |
| | | | (Estim) | |
Active(s) | Active | | (Actual) | | (percent) |
(sample #) |
====================================================================
====================================================================
==========
| | 1 | | | |
|
8 | +0 | 1 | 0 | | 8.33 | Cpu (2)
|
| -> 2 | HASH JOIN | | 400M | 2131 |
n s e
24 | +1 | 1 | 317M | 1138K | 91.67 | Cpu (22)
i ce
|
b l el
|
1 |
3 |
+8 |
INDEX FAST FULL SCAN | TEST_C_INDX |
1 | 20000 | |
20000 |
|
13 |
fer a
|
a n s
| -> 4 | INDEX FAST FULL SCAN | TEST_C_INDX |
o n -tr
20000 | 13 |
an
17 | +8 | 1 | 15872 | | |
|
h s
a deฺ
====================================================================
)
s Gui
====================================================================
e
r ฺ
mi dent
==========
p ฺ
g Stu
d
SQL> /
b i a@ this
i a (ja use
b to
A
i o
--------------------------------------------------------------------
n
nto--------------------------------------------------------------------
e A ----------------------------------------------------------------
J o s SQL Monitoring Report
SQL Text
--------------------------------------------------------------------
--------------------------------------------------------------------
-------------------------------------------
--------------------------------------------------------------------
--------------------------------------------------------------------
-------------------------------------------
Global Information
Status : DONE (ALL ROWS)
Instance ID : 1
Session ID : 124

-------------------------------------------------
| Elapsed | Cpu | Other | Fetch | Buffer |
| Time(s) | Time(s) | Waits(s) | Calls | Gets |
-------------------------------------------------
| 28 | 28 | 0.01 | 1 | 37 |
-------------------------------------------------

====================================================================
====================================================================
========
Detail |
n s e
| | | | (Estim) | |
ice
Active(s) | Active | | (Actual) | (Max) | (percent) |
b l el
(sample #) |
====================================================================
fe ra
====================================================================
a n s
========
n - tr
o
an
1 | +30 | 1 | 1 | | |
|
h s
a deฺ
| 1 | SORT AGGREGATE |
e )
s Gui | 1 | |
r ฺ
mi dent
31 | +0 | 1 | 1 | | 10.00 | Cpu (3)
|
p ฺ
g Stu
| 2 | HASH JOIN
d | | 400M | 2131 |
30 |
|
+1 |
b
1 |
i a@ this
400M | 1138K | 90.00 | Cpu (27)
| 3 |
i a (ja use
INDEX FAST FULL SCAN | TEST_C_INDX | 20000 | 13 |
1 |
Ab
+8 |
to
1 | 20000 | | |
o
|
n io
nt
e A 23 | +8 | 1 | 20000 | | |
J o s |
====================================================================
====================================================================
========
SQL>
13) When your query is executed, what do you observe in your session 1?
a) Session 1 also reports the same execution plan as the one you observed in session
2.
Session 1:
----------
COUNT(*)

----------
400000000
Elapsed: 00:00:30.70
Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
------------------
(%CPU)| Time |
--------------------------------------------------------------------
------------------
n s e
(99)| 00:00:22 |
i ce
b l el
|
|* 2 |
|
HASH JOIN | |
fe
400M| 2288M| 2131 r a
(99)| 00:00:22 |
a n s
|* 3 | - tr
n 13
o
an
(0)| 00:00:01 |
(0)| 00:00:01 |
h s
a deฺ
)
s Gui
--------------------------------------------------------------------
e
------------------ r ฺ
i ent
p ฺ m d operation id):
d g
Predicate Information (identified
S t uby
b i a@ this
---------------------------------------------------
i a (ja use
2 - access("T1"."C"="T2"."C")
b to
3 - filter("T1"."C"=1)
A
i o
4 - filter("T2"."C"=1)
nton
e A
J o s Statistics
----------------------------------------------------------
0 recursive calls
0 db block gets
90 consistent gets
0 physical reads
0 redo size
0 sorts (memory)
0 sorts (disk)
1 rows processed
SQL>
14) In session 1, disable autotrace.
Session 1:
----------

SQL>
15) From your session 1, how can you ensure that you gather all execution plan statistics
for the following query without changing any session parameters? Implement your
solution.
Session 1:
----------
SQL> @ep_execute_with_all
SQL> set echo on
SQL>
n s e
SQL> set timing on
ice
SQL>
select /*+ gather_plan_statistics */ count(*) from test t1, bl
el
SQL>
test t2 where t1.c=t2.c and t1.c=1; f e ra
t r a ns
COUNT(*)
o n -
an
----------
400000000
h s
a deฺ
Elapsed: 00:00:59.95
e )
s Gui
r ฺ
mi dent
SQL>
p
g Stuฺ
16) From your session 1, retrieve
d
@ hexecution
all plans corresponding to all the queries you
executed since the b i a t i s
( j abeginning
s eof this lab. What is your conclusion?
u
bia waytoto find out all the plans is to look at the content of the SGA using
a) The easiest
A
iothe dbms_xplan.display_cursor function. First, you must determine the
o n
t SQL_Ids used to represent your queries. You essentially have two queries, and
e An one that has two children. You should now understand what happened at step 4.
Jos The fact that there was no cost information is probably due to the use of the rule-
based optimizer instead of the cost-based one.
Session 1:
----------
SQL> @ep_retrieve_all_plans
SQL> set echo on
SQL>
SQL>
SQL> col sql_text format a50
SQL>
SQL> select sql_id,plan_hash_value,sql_text from v$sql where
sql_text like '%from test t1, test t2%';
SQL_ID PLAN_HASH_VALUE SQL_TEXT

------------- --------------- --------------------------------------
------------

dkz7v96ym42c6 3253233075 select count(*) from test t1, test t2
where t1.c=t
2.c and t1.c=1
dkz7v96ym42c6 1643938535 select count(*) from test t1, test t2

where t1.c=t
2.c and t1.c=1
8w580dd6ncgqw 3253233075 select /*+ gather_plan_statistics */

count(*) from
test t1, test t2 where t1.c=t2.c and
t1.c=1
0w0va2d7hhtxa 3253233075 explain plan for select count(*) from

test t1, tes
t t2 where t1.c=t2.c and t1.c=1
n s e
i ce
dd09kf5dnp1gt 903671040 select sql_id,plan_hash_value,sql_text
b l el
from v$sql
fer
where sql_text like '%from test t1, a
test t2%'
a n s
o n -tr
an
32fqwuk16uf23 3253233075 EXPLAIN PLAN SET
STATEMENT_ID='PLUS2140495' FOR se
h s
a deฺ
lect count(*) from test t1, test t2
where t1.c=t2.
e )
s Gui
r ฺ
mi dent
c and t1.c=1
p ฺ
g Stu
d
6 rows selected.
b i a@ this
i a (ja use
Elapsed: 00:00:00.02
SQL>
A b * from to
ni o
SQL> select
ntotable(dbms_xplan.display_cursor('dkz7v96ym42c6',null,'TYPICAL'));
e A
J o s PLAN_TABLE_OUTPUT
--------------------------------------------------------------------
--------------------------------------------------------------------
----------------------------------------------------------------
SQL_ID dkz7v96ym42c6, child number 0
-------------------------------------
--------------------------------------------------------------------
------------------
(%CPU)| Time |
--------------------------------------------------------------------
------------------
| 0 | SELECT STATEMENT | | | | 2131
(100)| |
| |

|* 2 | HASH JOIN | | 400M| 2288M| 2131
(99)| 00:00:22 |
(0)| 00:00:01 |
(0)| 00:00:01 |
--------------------------------------------------------------------
------------------

---------------------------------------------------
2 - access("T1"."C"="T2"."C")
3 - filter("T1"."C"=1)
4 - filter("T2"."C"=1)
n s e
SQL_ID dkz7v96ym42c6, child number 1
i ce
-------------------------------------
b l el
fer a
a n s
n -tr
o
an
------------------------------------------
------------------------------------------
h s
a deฺ
e )
s Gui
|
r ฺ
mi dent
| 2 | NESTED LOOPS |
p ฺ
g Stu
|
|* 3 |
d
INDEX RANGE SCAN| TEST_C_INDX |
|* 4 |
b i a@ this
INDEX RANGE SCAN| TEST_C_INDX |
------------------------------------------

b
Predicate
A to
Information
o
ni 3 - access("T1"."C"=1)
---------------------------------------------------
n t o
e A
Jos
4 - access("T1"."C"="T2"."C")
Note
-----
- rule based optimizer used (consider using cbo)
49 rows selected.
Elapsed: 00:00:00.04
SQL>
SQL> select * from
table(dbms_xplan.display_cursor('8w580dd6ncgqw',null,'ADVANCED
ALLSTATS LAST'));
PLAN_TABLE_OUTPUT
--------------------------------------------------------------------
--------------------------------------------------------------------
----------------------------------------------------------------
SQL_ID 8w580dd6ncgqw, child number 0
-------------------------------------

select /*+ gather_plan_statistics */ count(*) from test t1, test t2
where t1.c=t2.c and t1.c=1
--------------------------------------------------------------------
--------------------------------------------------------------------
-------------------
| Id | Operation | Name | Starts | E-Rows |E-
Bytes| Cost (%CPU)| E-Time | A-Rows | A-Time | Buffers | OMem
| 1Mem | Used-Mem |
--------------------------------------------------------------------
--------------------------------------------------------------------
-------------------
6 | | | 1 |00:00:59.94 | 90 | |
n s e
| |
i ce
|* 2 | HASH JOIN | | 1 | 400M|
b l el
2288M| 2131 (99)| 00:00:22 |
1155K| 1155K| 1115K (0)|
400M|00:00:00.01 | 90 |
fer a
|* 3 | INDEX FAST FULL SCAN| TEST_C_INDX | n s
1 | 20000 |
a
60000 | 13 (0)| 00:00:01 | 20000 |00:00:00.02 |
o n -tr 45 |
an
| | |
|* 4 | INDEX FAST FULL SCAN| TEST_C_INDX | 1 | 20000 |
60000 | 13
h s
a deฺ
(0)| 00:00:01 | 20000 |00:00:00.16 | 45 |
| | |
e )
s Gui
r ฺ
mi dent
--------------------------------------------------------------------
p ฺ
--------------------------------------------------------------------
g Stu
-------------------
d
@ Alias
Query Block Namebi/aObjectth is (identified by operation id):
i a (ja use
-------------------------------------------------------------
A b to
i o
1 - SEL$1
nton 34 -- SEL$1 / T1@SEL$1
e A SEL$1 / T2@SEL$1
J o s Outline Data
-------------
/*+
BEGIN_OUTLINE_DATA
IGNORE_OPTIM_EMBEDDED_HINTS
OPTIMIZER_FEATURES_ENABLE('11.1.0.6')
DB_VERSION('11.1.0.6')
ALL_ROWS
OUTLINE_LEAF(@"SEL$1")
INDEX_FFS(@"SEL$1" "T1"@"SEL$1" ("TEST"."C"))
INDEX_FFS(@"SEL$1" "T2"@"SEL$1" ("TEST"."C"))
LEADING(@"SEL$1" "T1"@"SEL$1" "T2"@"SEL$1")
USE_HASH(@"SEL$1" "T2"@"SEL$1")
END_OUTLINE_DATA
*/

---------------------------------------------------

2 - access("T1"."C"="T2"."C")
3 - filter("T1"."C"=1)
4 - filter("T2"."C"=1)

-----------------------------------------------------------
1 - (#keys=0) COUNT(*)[22]
2 - (#keys=1)
3 - "T1"."C"[NUMBER,22]
4 - "T2"."C"[NUMBER,22]
55 rows selected.
Elapsed: 00:00:00.12
n s e
SQL>
i ce
b l el
--------------------------------------------------------------
fe r a
set echo on
a n s
o n -tr
an
col sql_text format a50

h s
a deฺ
e )
s from i where sql_text
uv$sql
r ฺ G
mi dent
select sql_id,plan_hash_value,sql_text
p ฺ
like '%from test t1, test t2%';
g Stu
d
select * from
b i a@ this
table(dbms_xplan.display_cursor('dkz7v96ym42c6',null,'TYPICAL'));
i a (ja use
b
select *
A
from
to
ni o
table(dbms_xplan.display_cursor('8w580dd6ncgqw',null,'ADVANCED
ntoALLSTATS LAST'));
e A
J o s 17) From session 1, try to retrieve your execution plans from the Automatic Workload
Repository. What happens and why?
a) You can use the previously found SQL_Ids to search through the
DBA_HIST_SQLTEXT view. You should see that right now, none of your
queries were stored in the AWR.
Session 1:
----------
SQL> @ep_retrieve_awr
SQL> set echo on
SQL>
SQL>
SQL> SELECT SQL_ID, SQL_TEXT FROM dba_hist_sqltext
2 WHERE SQL_ID in ('dkz7v96ym42c6','8w580dd6ncgqw');
no rows selected

SQL>
Elapsed: 00:00:00.01
SQL>
--------------------------------------------------------------
set echo on
set linesize 200
SELECT SQL_ID, SQL_TEXT FROM dba_hist_sqltext

WHERE SQL_ID in ('dkz7v96ym42c6','8w580dd6ncgqw');
n s e
18) How can you ensure that you retrieve your queries from the Automatic Workload
i ce
Repository? Implement your solution.
b l el
a) You must flush the SGA information to the AWR. You can use fer a
DBMS_WORKLOAD_REPOSITORY.CREATE_SNAPSHOT for this purpose. a n s
o n -tr
Session 1: a n
----------
h s
a deฺ
e )
s Gui
r ฺ
mi dent
SQL> @ep_save_awr
SQL> set echo on
p ฺ
g Stu
SQL>
d
b i a@ this
SQL> EXEC DBMS_WORKLOAD_REPOSITORY.CREATE_SNAPSHOT('ALL');
(ja successfully
PL/SQL procedure
i a u se completed.
A to
b 00:00:02.27
ni o
Elapsed:
ntoSQL>
e A
J o s --------------------------------------------------------------
set echo on
EXEC DBMS_WORKLOAD_REPOSITORY.CREATE_SNAPSHOT('ALL');
19) Verify that your solution works.
Session 1:
----------
SQL> @ep_show_awr
SQL> set echo on
SQL>
SQL>
SQL> SELECT PLAN_TABLE_OUTPUT
2 FROM
3 TABLE (DBMS_XPLAN.DISPLAY_AWR('dkz7v96ym42c6'));

PLAN_TABLE_OUTPUT
--------------------------------------------------------------------
--------------------------------------------------------------------
----------------------------------------------------------------
SQL_ID dkz7v96ym42c6
--------------------
------------------------------------------
------------------------------------------
n s e
| 2 | NESTED LOOPS | |
i ce
| 3 | INDEX RANGE SCAN| TEST_C_INDX |
b l el
| 4 | INDEX RANGE SCAN| TEST_C_INDX |
------------------------------------------
fer a
a n s
Note
o n -tr
an
-----
- rule based optimizer used (consider using cbo)
h s
a deฺ
SQL_ID dkz7v96ym42c6
e )
s Gui
-------------------- r ฺ
i t2enwheret t1.c=t2.c and t1.c=1
ฺ
select count(*) from test t1,
p mtest d
d g Stu
b i a@ this
i a (ja use
--------------------------------------------------------------------
b to
------------------
| Id A
n i o | Operation | Name | Rows | Bytes | Cost
nto--------------------------------------------------------------------
(%CPU)| Time |
e A
J o s ------------------
(100)| |
| |
| 2 | HASH JOIN | | 100M| 572M| 543
(98)| 00:00:06 |
| 3 | INDEX FAST FULL SCAN| TEST_C_INDX | 10000 | 30000 | 8
(0)| 00:00:01 |
| 4 | INDEX FAST FULL SCAN| TEST_C_INDX | 10000 | 30000 | 8
(0)| 00:00:01 |
--------------------------------------------------------------------
------------------
36 rows selected.
Elapsed: 00:00:00.04
SQL>
SQL> SELECT PLAN_TABLE_OUTPUT
2 FROM

3 TABLE
(DBMS_XPLAN.DISPLAY_AWR('8w580dd6ncgqw',null,null,'TYPICAL ALLSTATS
LAST'));
PLAN_TABLE_OUTPUT
--------------------------------------------------------------------
--------------------------------------------------------------------
----------------------------------------------------------------
SQL_ID 8w580dd6ncgqw
--------------------
select /*+ gather_plan_statistics */ count(*) from test t1, test t2
where t1.c=t2.c and t1.c=1
--------------------------------------------------------------------
n s e
-------------------
i ce
| Id | Operation | Name | E-Rows |E-Bytes| Cost
b l el
(%CPU)| E-Time |
fer
-------------------------------------------------------------------- a
-------------------
a n s
n -
| tr |
o
an
2131 (100)| |
| |
h s
a deฺ
| 2 | HASH JOIN |
e )
s Gui | 400M| 2288M|
r ฺ
mi dent
2131 (99)| 00:00:22 |
| 3 | ฺ
p
g Stu
13 (0)| 00:00:01 |
d
|
13
4 |
b
(0)| 00:00:01 |i a@ this
i a (ja use
--------------------------------------------------------------------
b to
-------------------
A
o n io
Note
nt
e A -----
J o s - Warning: basic plan statistics not available. These are only

collected when:
* hint 'gather_plan_statistics' is used for the statement or
* parameter 'statistics_level' is set to 'ALL', at session or
system level
23 rows selected.
Elapsed: 00:00:00.02
SQL>
--------------------------------------------------------------
set echo on
SELECT PLAN_TABLE_OUTPUT
FROM
TABLE (DBMS_XPLAN.DISPLAY_AWR('dkz7v96ym42c6'));

SELECT PLAN_TABLE_OUTPUT
FROM
TABLE (DBMS_XPLAN.DISPLAY_AWR('8w580dd6ncgqw',null,null,'TYPICAL
ALLSTATS LAST'));
20) Exit from both SQL*Plus sessions, and clean up your environment by executing the
ep_cleanup.sh script from one of your terminal sessions.
Session 1:
----------
SQL> exit
n s e
i ce
b l el
[oracle@edrsr33p1-orcl Explain_Plan]$ ./ep_cleanup.sh fe r a
a n s
n -tr
o
n
a reserved.
h s
a deฺ
e )
s Gui
r ฺ
miEdition t
Connected to:
p
Oracle Database 11g Enterprise ฺ d e nRelease 11.1.0.6.0 -
Production g
dOracle S t u
@ s
thi options
ia Testing
a b
j e
SQL> b ia ( to us
i
SQL> A user ep cascade;
o drop
to n
A n User dropped.
o s e
J SQL>
SQL> exit;
--------------------------------------------------------------
#!/bin/bash

export
sqlplus / as sysdba @ep_cleanup.sql

--------------------------------------------------------------
set echo on
drop user ep cascade;
exit;
Do not forget to exit from your session 2:

n s e
i ce
Session 2: b l el
----------
fer a
a n s
SQL> exit
n - tr
o
s an
) h a deฺ
r ฺ
[oracle@edrsr33p1-orcl Explain_Plan]$e s Gui
p ฺ mi dent
d g Stu
b i a@ this
i a (ja use
Ab to
o n io
nt
e A
J o s

n s e
i ce
b l el
fer a
a n s
n - tr
o
s an
) h a deฺ
r ฺ e s Gui
p ฺ mi dent
d g Stu
b i a@ this
i a (ja use
Ab to
o n io
nt
e A
J o s

Practice 6-1: Star Schema Tuning
In this practice, you optimize a query to use star transformation and access the benefits of
using this optimizer technique.
1) From a terminal session, connected as the oracle user, execute the
setup_star_schema_lab.sh script located in your
/home/oracle/solutions/Star_Schema_Tuning directory.
[oracle@edrsr33p1-orcl Star_Schema_Tuning]$
./setup_star_schema_lab.sh
SQL*Plus: Release 11.1.0.6.0 - Production on Fri Mar 21 00:17:20

2008
n s e
ice
b l el
fe ra
Connected to:
a n s
-tr
Production
o n
an
h s
a deฺ
SQL> SQL> SQL> SQL>
e )
s Gui
r ฺ
mi dent
Grant succeeded.
p ฺ
g Stu
SQL> SQL> d
User altered.
b i a@ this
i a (ja usefrom Oracle Database 11g Enterprise Edition
SQL> SQL> Disconnected
to - Production
Release b11.1.0.6.0
A
n o
i Real Application Testing
nt oand options
e A [oracle@edrsr33p1-orcl Star_Schema_Tuning]$
J o s --------------------------------------------------------------
#!/bin/bash
cd /home/oracle/solutions/Star_Schema_Tuning
export
sqlplus / as sysdba <<FIN!
set echo on
grant dba to sh;

Practice 6-1: Star Schema Tuning (continued)
FIN!
2) From the same terminal window, start a SQL*Plus session connected as the SH user
and do not disconnect from it until this lab finishes. Before executing the following
SQL statement, ensure that you flush both the shared pool and the buffer cache to
avoid caching issues as much as possible. After this, analyze the execution of the
following query (You can use the query.sql script.):

FROM sales s, times t, customers c, channels ch
n s e
ce
el i
a b l
ch.channel_desc in ('Internet','Catalog') AND
fe r
a n s
t.calendar_quarter_desc IN ('1999-01','1999-02','2000-03','2000-04')
-tr
o n
What are your conclusions?
s an
a) As you can see in the output of the executionhplan,
) i eฺ seems to use a large
a thisdquery
number of bytes to access the SALESฺe
r s Basically,
table. G u the optimizer performs a
i t
enbest way to handle it.
ฺmnot ubedthe
full scan of this table. This might
d g p t
@ s S
j a b e thi
ia Star_Schema_Tuning]$ sqlplus sh/sh
SQL*Plus:ia
( u s
2008 A b t o
Release 11.1.0.6.0 - Production on Fri Mar 21 00:31:10
o n io
nt Copyright (c) 1982, 2007, Oracle. All rights reserved.
e A
J o s
Connected to:
Production
SQL> @first_run
SQL>
System altered.
System altered.
SQL>

SQL> set timing on
SQL>
2 SUM(s.amount_sold) sales_amount
3 FROM sales s, times t, customers c, channels ch
4 WHERE s.time_id = t.time_id

5 AND s.cust_id = c.cust_id
6 AND s.channel_id = ch.channel_id
7 AND c.cust_state_province = 'CA'
8 AND ch.channel_desc in ('Internet','Catalog')
9 AND t.calendar_quarter_desc IN ('1999-01','1999-02','2000-03','2000-04')
10 GROUP BY ch.channel_class, c.cust_city, t.calendar_quarter_desc;
CHANNEL_CLASS CUST_CITY CALENDA

SALES_AMOUNT
-------------------- ------------------------------ ------- --------
n s e
----
i ce
Indirect
987.3
Quartzhill 1999-01
b l el
Indirect Arbuckle 1999-02
fe r a
…
a n s
Montara 1999-02
n tr
-1999-02 1618.01
Indirect Quartzhill o
412.83
s an
) h a deฺ
41 rows selected.
r ฺ e s Gui
Elapsed: 00:00:00.68
p ฺ mi dent
d g Stu
a@ this
Execution Plan
i
----------------------------------------------------------
b
(ja use
b i a
-------------------------------------------------------------------------------------------------------------
| Id | Operation
to | Name | Rows | Bytes | Cost (%CPU)| Time | Pstart| Pstop |
|
i A
-------------------------------------------------------------------------------------------------------------
o0 | SELECT STATEMENT | | 1144 | 96096 | 935 (4)| 00:00:12 | | |
ton
| 1 | HASH GROUP BY | | 1144 | 96096 | 935 (4)| 00:00:12 | | |
|* 2 | HASH JOIN | | 6231 | 511K| 934 (3)| 00:00:12 | | |
A n |* 3 |
|* 4 |
TABLE ACCESS FULL
HASH JOIN
| CHANNELS |
|
2 |
| 12462 |
42 |
766K|
3
930
(0)| 00:00:01 |
(3)| 00:00:12 |
|
|
|
|
o s e | 5 | PART JOIN FILTER CREATE | :BF0000 | 365 | 5840 | 18 (0)| 00:00:01 | | |
J
|* 6 | TABLE ACCESS FULL | TIMES | 365 | 5840 | 18 (0)| 00:00:01 | | |
|* 7 | HASH JOIN | | 49822 | 2286K| 911 (3)| 00:00:11 | | |
|* 8 | TABLE ACCESS FULL | CUSTOMERS | 383 | 9958 | 406 (1)| 00:00:05 | | |
| 9 | PARTITION RANGE JOIN-FILTER| | 918K| 18M| 498 (4)| 00:00:06 |:BF0000|:BF0000|
| 10 | TABLE ACCESS FULL | SALES | 918K| 18M| 498 (4)| 00:00:06 |:BF0000|:BF0000|
-------------------------------------------------------------------------------------------------------------

---------------------------------------------------
2 -
access("S"."CHANNEL_ID"="CH"."CHANNEL_ID")
3 -
filter("CH"."CHANNEL_DESC"='Catalog' OR "CH"."CHANNEL_DESC"='Internet')
4 -
access("S"."TIME_ID"="T"."TIME_ID")
6 -
filter("T"."CALENDAR_QUARTER_DESC"='1999-01' OR "T"."CALENDAR_QUARTER_DESC"='1999-02' OR
"T"."CALENDAR_QUARTER_DESC"='2000-03' OR "T"."CALENDAR_QUARTER_DESC"='2000-04')
Statistics
----------------------------------------------------------
0 db block gets
2021 physical reads
0 redo size

141 sorts (memory)
0 sorts (disk)
41 rows processed
SQL>
SQL>
--------------------------------------------------------------
set echo on

set pagesize 200

n s e
set linesize 250
i ce
set timing on
b l el
set autotrace on
fe r a
a n s
n - tr
o
an
WHERE s.time_id = t.time_id
AND s.cust_id = c.cust_id
h s
a deฺ
AND s.channel_id = ch.channel_id
e )
s Gui
r ฺ
mi dent
AND c.cust_state_province = 'CA'
AND ฺ
ch.channel_desc in ('Internet','Catalog')
p
g Stu
AND
d
t.calendar_quarter_desc IN ('1999-01','1999-02','2000-
03','2000-04')
b i a@ this
i a (ja use
Ab to
3) Without modifying the SH schema, how can you improve the execution plan for the
o n io
query mentioned in step 2? Verify your solution and explain why it is probably a
nt
e A better solution.
J o s a) Enable star transformation in your session. In this step, you do not want to use a
temporary table for the star transformation. Looking at the previous execution
plan, the optimizer estimates the data that is to be manipulated in megabytes.
Using the star transformation as follows, the estimation is now expressed in
kilobytes. That is why this new execution plan is probably a much better
alternative. However, note that this time the CUSTOMERS table is accessed using
full scan twice. If the table is larger, the impact is significant.
SQL> @second_run
SQL> set echo on
SQL>
SQL>
System altered.
Elapsed: 00:00:00.10

System altered.
Elapsed: 00:00:00.20
SQL>
SQL> set timing on

SQL>
SQL>
SQL> ALTER SESSION SET star_transformation_enabled=TEMP_DISABLE;
Session altered.
Elapsed: 00:00:00.00
SQL>
n s e
SQL>
i ce
b l el
fer a
a n s
o n -tr
an
8 AND
h s
a deฺ
9 AND )
s Gui
e
r ฺ
mi dent
03','2000-04')
p ฺ
g Stu
d
@ his
CHANNEL_CLASS bia CUST_CITY
t CALENDA
SALES_AMOUNT(ja e
b ia to us------------------------------ ------- --------
--------------------
---- A
o n io
Indirect San Francisco 2000-04
nt 13227.99
e A
J o s Indirect
1319
Montara 2000-04
Indirect Cloverdale 2000-04

7.27
…
Indirect Pala 2000-03
Indirect Montara 1999-02
1618.01
Indirect Quartzhill 1999-02
412.83
3058.27
3263.93
41 rows selected.
Elapsed: 00:00:00.33
Execution Plan
----------------------------------------------------------

------------------------------------------------------------------------------------------------------------------
----------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
Pstart| Pstop |
------------------------------------------------------------------------------------------------------------------
----------
| 0 | SELECT STATEMENT | | 3 | 252 | 983 (1)| 00:00:12 |
| |
| 1 | HASH GROUP BY | | 3 | 252 | 983 (1)| 00:00:12 |
| |
|* 2 | HASH JOIN | | 3 | 252 | 982 (1)| 00:00:12 |

| |
|* 3 | HASH JOIN | | 7 | 441 | 978 (1)| 00:00:12 |
| |
|* 4 | HASH JOIN | | 27 | 1269 | 960 (1)| 00:00:12 |
| |
|* 5 | TABLE ACCESS FULL | CUSTOMERS | 383 | 9958 | 406 (1)| 00:00:05 |
| |
| 6 | PARTITION RANGE SUBQUERY | | 507 | 10647 | 553 (1)| 00:00:07
|KEY(SQ)|KEY(SQ)|
| 7 | TABLE ACCESS BY LOCAL INDEX ROWID| SALES | 507 | 10647 | 553 (1)| 00:00:07
|KEY(SQ)|KEY(SQ)|
| 8 | BITMAP CONVERSION TO ROWIDS | | | | | |
| |
| 9 | BITMAP AND | | | | | |
|
| 10 |
|
BITMAP MERGE | | | | | |
n s e
| |
ice
el
| 11 | BITMAP KEY ITERATION | | | | | |
|
| 12 |
|
BUFFER SORT | | | | |
a b|l
|
|* 13 |
|
TABLE ACCESS FULL | CHANNELS | 2 | 42 | 3
fe r
(0)| 00:00:01 |
| |
a n s
tr
|* 14 | BITMAP INDEX RANGE SCAN | SALES_CHANNEL_BIX | | | |
|KEY(SQ)|KEY(SQ)|
| 15 | BITMAP MERGE |
o n - | | | | |
an
| |
|
| 17 |
|
BUFFER SORT
h
| s
a deฺ | | | | |
|
|* 18 |
|
TABLE ACCESS FULL
e )
s Gui
| TIMES | 365 | 5840 | 18 (0)| 00:00:01 |
| |
r ฺ
mi dent
|* 19 | BITMAP INDEX RANGE SCAN | SALES_TIME_BIX | | | |
|KEY(SQ)|KEY(SQ)|
| 20 | BITMAP MERGE
p ฺ
g Stu
| | | | | |
| |
d
a@ this
|
| 22 |
|
b i
BUFFER SORT | | | | | |
(ja use
| |
a
| |
b i to
|* 24 |
io A
|KEY(SQ)|KEY(SQ)|
BITMAP INDEX RANGE SCAN | SALES_CUST_BIX | | | |
t o|n
|* 25 |
|
TABLE ACCESS FULL | TIMES | 365 | 5840 | 18 (0)| 00:00:01 |
A n |* 26 |
| |
TABLE ACCESS FULL | CHANNELS | 2 | 42 | 3 (0)| 00:00:01 |
o s e ------------------------------------------------------------------------------------------------------------------
----------
J Predicate Information (identified by operation id):

---------------------------------------------------
13 - filter("CH"."CHANNEL_DESC"='Catalog' OR "CH"."CHANNEL_DESC"='Internet')
18 - filter("T"."CALENDAR_QUARTER_DESC"='1999-01' OR "T"."CALENDAR_QUARTER_DESC"='1999-02' OR
Note
-----
- star transformation used for this statement
Statistics
----------------------------------------------------------
0 db block gets

2130 physical reads
0 redo size
147 sorts (memory)
0 sorts (disk)
41 rows processed
SQL>
--------------------------------------------------------------
set echo on
n s e
ice
b l el
set pagesize 200
fe ra
set linesize 250
a n s
set timing on
o n -tr
an
set autotrace on
h s
a deฺ
)
s Gui
ALTER SESSION SET star_transformation_enabled=TEMP_DISABLE;
e
r ฺ
i ent
p ฺ m
SELECT ch.channel_class, gc.cust_city, d t.calendar_quarter_desc,
d S t u
SUM(s.amount_sold)
b
FROM sales s, timesi a@ t h is c, channels ch
sales_amount
t, customers
i a (ja == t.time_id
WHERE s.time_id
u se
AND
AND A
b
s.cust_id
to = ch.channel_id
c.cust_id
n o
i c.cust_state_province = 'CA'
s.channel_id
ntoAND
e A AND ch.channel_desc in ('Internet','Catalog')
J o s AND t.calendar_quarter_desc IN ('1999-01','1999-02','2000-

03','2000-04')
4) How would you enhance the previous optimization without changing the SH schema?
a) Let the optimizer decide if it is better to use a temporary table. You can try to set
the STAR_TRANSFORMATION_ENABLED parameter to TRUE.
SQL> @third_run
SQL> set echo on
SQL>
System altered.
Elapsed: 00:00:00.10

System altered.
Elapsed: 00:00:00.21
SQL>
SQL> set timing on

SQL>
SQL> ALTER SESSION SET star_transformation_enabled=TRUE;
Session altered.
Elapsed: 00:00:00.00
SQL>
n s e
ice
b l el
fe r a
a n s
o n -tr
an
9 AND t.calendar_quarter_desc IN ('1999-01','1999-02','2000-
03','2000-04')
h s
a deฺ
e )
s Gui
t.calendar_quarter_desc; r ฺ
i ent
p ฺ m d
CHANNEL_CLASS
d g
CUST_CITY
S t u CALENDA
SALES_AMOUNT
b
--------------------i a@------------------------------
t h is ------- --------
j a e
ia ( to usSan Francisco
----
Indirect
A b 2000-04
o io
13227.99
n
nt 1319
e A
J o s …
3058.27
3263.93
41 rows selected.
Elapsed: 00:00:00.30
Execution Plan
----------------------------------------------------------
------------------------------------------------------------------------------------------------------------------
------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time
| Pstart| Pstop |
------------------------------------------------------------------------------------------------------------------
------------------
| 0 | SELECT STATEMENT | | 3 | 219 | 985 (1)| 00:00:12
| | |
| 1 | TEMP TABLE TRANSFORMATION | | | | |
| | |
| 2 | LOAD AS SELECT | SYS_TEMP_0FD9D660F_8947E | | | |
| | |
|* 3 | TABLE ACCESS FULL | CUSTOMERS | 383 | 9958 | 406 (1)| 00:00:05
| | |

| 4 | HASH GROUP BY | | 3 | 219 | 579 (2)| 00:00:07
| | |
|* 5 | HASH JOIN | | 3 | 219 | 578 (1)| 00:00:07
| | |
|* 6 | HASH JOIN | | 7 | 364 | 574 (1)| 00:00:07
| | |
|* 7 | HASH JOIN | | 28 | 1008 | 556 (1)| 00:00:07
| | |
| 8 | TABLE ACCESS FULL | SYS_TEMP_0FD9D660F_8947E | 383 | 5745 | 2 (0)| 00:00:01
| | |
| 9 | PARTITION RANGE SUBQUERY | | 507 | 10647 | 553 (1)| 00:00:07

|KEY(SQ)|KEY(SQ)|
| 10 | TABLE ACCESS BY LOCAL INDEX ROWID| SALES | 507 | 10647 | 553 (1)| 00:00:07
|KEY(SQ)|KEY(SQ)|
| 11 | BITMAP CONVERSION TO ROWIDS | | | | |
| | |
| 12 | BITMAP AND | | | | |
| | |
| 13 | BITMAP MERGE | | | | |
| | |
| 14 | BITMAP KEY ITERATION | | | | |
| | |
| 15 | BUFFER SORT | | | | |
| | |
|* 16 | TABLE ACCESS FULL | CHANNELS | 2 | 42 | 3 (0)| 00:00:01
|
|* 17 |
| |
BITMAP INDEX RANGE SCAN | SALES_CHANNEL_BIX | | | |
n s e
|KEY(SQ)|KEY(SQ)|
ice
el
| 18 | BITMAP MERGE | | | | |
|
| 19 |
| |
BITMAP KEY ITERATION | | | |
a
|
b l
|
| 20 |
| |
BUFFER SORT | | | |
fe r |
| | |
a n s
tr
|* 21 | TABLE ACCESS FULL | TIMES | 365 | 5840 | 18 (0)| 00:00:01
|
|* 22 |
| |
BITMAP INDEX RANGE SCAN
n -
| SALES_TIME_BIX
o
| | | |
an
|KEY(SQ)|KEY(SQ)|
| 23 | BITMAP MERGE | | | | |
|
| 24 |
| |
BITMAP KEY ITERATION
h
| s
a deฺ | | | |
|
| 25 |
| |
BUFFER SORT
e )
s Gui
| | | | |
| | |
r ฺ
mi dent
| 26 | TABLE ACCESS FULL | SYS_TEMP_0FD9D660F_8947E | 1 | 13 | 2 (0)| 00:00:01
|
|* 27 |
| |
p ฺ
g Stu
BITMAP INDEX RANGE SCAN | SALES_CUST_BIX | | | |
|KEY(SQ)|KEY(SQ)|
d
a@ this
|* 28 | TABLE ACCESS FULL | TIMES | 365 | 5840 | 18 (0)| 00:00:01
|
|* 29 |
| |
b i
TABLE ACCESS FULL | CHANNELS | 2 | 42 | 3 (0)| 00:00:01
(ja use
| | |
a
------------------------------------------------------------------------------------------------------------------
b i
------------------
to
io A
t o n
---------------------------------------------------
n 3 - filter("C"."CUST_STATE_PROVINCE"='CA')
eA
Jos
7 - access("S"."CUST_ID"="C0")
27 - access("S"."CUST_ID"="C0")
Note
-----
Statistics
----------------------------------------------------------
19 db block gets
2157 physical reads
1616 redo size

155 sorts (memory)
0 sorts (disk)
41 rows processed
SQL>
SQL>
--------------------------------------------------------------
set echo on

set pagesize 200

n s e
set linesize 250
i ce
set timing on
b l el
set autotrace on
fe r a
ALTER SESSION SET star_transformation_enabled=TRUE;
a n s
n - tr
o
an
h s
a deฺ
e )
s Gui
r ฺ
mi dent
AND ฺ
p
g Stu
AND
d
c.cust_state_province = 'CA'
AND
AND b i a@ this
03','2000-04')
i a (ja use
b to
A
o n io
t
n5) How do you eliminate one table access on the CUSTOMERS table from the previous
e A
J o s execution plan for the same SELECT statement seen in step 3?
a) Create a bitmap join index between the SALES and CUSTOMERS tables.
6) Try to apply your finding. What happens and why?
a) Because the CUSTOMERS_PK primary key constraint is not enforced, it is not
possible to create a bitmap join index between the SALES and CUSTOMERS
tables.
SQL> @create_bji.sql
SQL>
SQL> CREATE BITMAP INDEX sales_c_state_bjix ON
sales(customers.cust_state_province) FROM sales, customers WHERE
sales.cust_id=customers.cust_id
2 LOCAL NOLOGGING COMPUTE STATISTICS;
CREATE BITMAP INDEX sales_c_state_bjix ON

*
ERROR at line 1:
ORA-25954: missing primary key or unique constraint on dimension
SQL>
--------------------------------------------------------------
set echo on

LOCAL NOLOGGING COMPUTE STATISTICS;
n s e
i ce
7) Fix the issue you found and apply your solution from step 5 again. b l el
fer a
a) You need to ENABLE VALIDATE the CUSTOMERS_PK constraint before you
a n s
can create the bitmap join index.
o n -tr
s an
SQL> @recreate_bji
) h a deฺ
SQL>
SQL> alter table customers enableฺe
i
s Gucustomers_pk;
r
mi dent
constraint
Table altered. p ฺ
g Stu
d
SQL>
b i a@ this
SQL> CREATE (
a ja uINDEX
BITMAP se sales_c_state_bjix ON
b i to
o A
i
nton 2 LOCAL NOLOGGING COMPUTE STATISTICS;
e A Index created.
J o s
SQL>
SQL>
--------------------------------------------------------------
set echo on
alter table customers enable constraint customers_pk;

LOCAL NOLOGGING COMPUTE STATISTICS;
8) Verify that you solved the problem from step 5.
SQL> @fourth_run

SQL> set echo on
SQL>
SQL> ALTER SESSION SET star_transformation_enabled=TRUE;
Session altered.
SQL>
System altered.
System altered.
SQL>
n s e
ice
b l el
SQL> set timing on
fe r a
SQL>
a n s
-
n tr
o
an
h s
a deฺ
e )
s Gui
r ฺ
mi dent
7 AND ฺ
c.cust_state_province = 'CA'
p
g Stu
8 AND
d
9 AND
03','2000-04') b i a@ this
i a (ja use
b to
A
o n io
CHANNEL_CLASS CUST_CITY CALENDA
t
e An SALES_AMOUNT
Jos
-------------------- ------------------------------ ------- --------
----
987.3
Indirect Arbuckle 1999-02
241.2
…
1618.01
412.83
41 rows selected.
Elapsed: 00:00:00.23
Execution Plan
----------------------------------------------------------
------------------------------------------------------------------------------------------------------------------
-----------

| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
Pstart| Pstop |
------------------------------------------------------------------------------------------------------------------
-----------
| 0 | SELECT STATEMENT | | 498 | 41832 | 557 (2)| 00:00:07 |
| |
| 1 | HASH GROUP BY | | 498 | 41832 | 557 (2)| 00:00:07 |
| |
|* 2 | HASH JOIN | | 498 | 41832 | 556 (2)| 00:00:07 |
| |
|* 3 | TABLE ACCESS FULL | CHANNELS | 2 | 42 | 3 (0)| 00:00:01 |

| |
|* 4 | HASH JOIN | | 996 | 62748 | 552 (1)| 00:00:07 |
| |
|* 5 | TABLE ACCESS FULL | TIMES | 365 | 5840 | 18 (0)| 00:00:01 |
| |
|* 6 | HASH JOIN | | 3984 | 182K| 533 (1)| 00:00:07 |
| |
| |
| 8 | PARTITION RANGE SUBQUERY | | 73467 | 1506K| 126 (1)| 00:00:02
|KEY(SQ)|KEY(SQ)|
| 9 | TABLE ACCESS BY LOCAL INDEX ROWID| SALES | 73467 | 1506K| 126 (1)| 00:00:02
|KEY(SQ)|KEY(SQ)|
| 10 | BITMAP CONVERSION TO ROWIDS | | | | | |
|
| 11 |
|
BITMAP AND | | | | | |
n s e
| |
ice
el
| 12 | BITMAP MERGE | | | | | |
|
| 13 |
|
BITMAP KEY ITERATION | | | | |
a b l
|
|
| 14 |
|
BUFFER SORT | | | | |
fe r |
| |
a n s
tr
|* 15 | TABLE ACCESS FULL | CHANNELS | 2 | 42 | 3 (0)| 00:00:01 |
|
|* 16 |
|
BITMAP INDEX RANGE SCAN
n -
| SALES_CHANNEL_BIX |
o
| | |
an
|KEY(SQ)|KEY(SQ)|
| 17 | BITMAP MERGE | | | | | |
|
| 18 |
|
BITMAP KEY ITERATION
h
| s
a deฺ | | | | |
|
| 19 |
|
BUFFER SORT
e )
s Gui
| | | | | |
| |
r ฺ
mi dent
|* 20 | TABLE ACCESS FULL | TIMES | 365 | 5840 | 18 (0)| 00:00:01 |
|
|* 21 |
|
p ฺ
g Stu
BITMAP INDEX RANGE SCAN | SALES_TIME_BIX | | | |
|KEY(SQ)|KEY(SQ)|
d
a@ this
|* 22 | BITMAP INDEX SINGLE VALUE | SALES_C_STATE_BJIX | | | |
|KEY(SQ)|KEY(SQ)|
b i
------------------------------------------------------------------------------------------------------------------
(ja use
-----------
b i a to
A
---------------------------------------------------
io
t o n2
3
-
-
A n 4
5
-
-
access("S"."TIME_ID"="T"."TIME_ID")
filter("T"."CALENDAR_QUARTER_DESC"='1999-01' OR "T"."CALENDAR_QUARTER_DESC"='1999-02' OR
o s e 6 -
access("S"."CUST_ID"="C"."CUST_ID")
J 7
15
16
-
-
-
filter("C"."CUST_STATE_PROVINCE"='CA')
22 - access("S"."SYS_NC00008$"='CA')
Note
-----
Statistics
----------------------------------------------------------
0 db block gets
2002 physical reads
0 redo size
154 sorts (memory)
0 sorts (disk)

41 rows processed
SQL>
--------------------------------------------------------------
set echo on
ALTER SESSION SET star_transformation_enabled=TRUE;

set pagesize 200

set linesize 250
set timing on
n s e
set autotrace on
i ce
b l el
fer a
a n s
o n -tr
an
h s
a deฺ
AND )
s Gui
e
r ฺ
mi dent
AND t.calendar_quarter_desc IN ('1999-01','1999-02','2000-
03','2000-04')
p ฺ
g Stu
d
b i a@ this
9) Determine how(the
a ja system
u e dynamically determine which SALES partitions to
scould
i to query.
access forbthe previous
o A
t o ni View the OTHER column of PLAN_TABLE for the PARTITION RANGE
a)
An operation.
e
Jos SQL>
SQL>
@dynamic_partition_pruning

SQL> set timing off
SQL>
SQL> set long 4000
SQL>
SQL> alter session set star_transformation_enabled=true;
Session altered.
SQL>
2 SELECT ch.channel_class, c.cust_city, t.calendar_quarter_desc,

10 AND t.calendar_quarter_desc IN ('1999-01','1999-02','2000-
03','2000-04')
Explained.
SQL>
SQL> SELECT other
2 FROM plan_table
3 WHERE operation='PARTITION RANGE';
OTHER
--------------------------------------------------------------------
------------
n s e
SELECT distinct TBL$OR$IDX$PART$NUM("SALES", 0, d#, p#,
i ce
PAP_ALIAS_0."PAP_COL_ALI
b l el
AS_0") FROM (SELECT /*+ SEMIJOIN_DRIVER */ "T"."TIME_ID"
"PAP_COL_ALIAS_0" FROM
fer a
n
"TIMES" "T" WHERE "T"."CALENDAR_QUARTER_DESC"='1999-01' OR
a s
"T"."CALENDAR_QUARTER
o n -tr
an
_DESC"='1999-02' OR "T"."CALENDAR_QUARTER_DESC"='2000-03' OR
"T"."CALENDAR_QUART
h
ER_DESC"='2000-04') PAP_ALIAS_0 ORDER BY 1 s
a deฺ
e )
s Gui
r ฺ
SQL>
p ฺ mi dent
d g Stu
b i a@ this
--------------------------------------------------------------
i a (ja use
b
set echo
A
on
to
n o
i autotrace off
ntoset
e A set timing off
J o s set long 4000
alter session set star_transformation_enabled=true;
explain plan for

AND ch.channel_desc in ('Internet','Catalog')
AND t.calendar_quarter_desc IN ('1999-01','1999-02','2000-
03','2000-04')
SELECT other
FROM plan_table
WHERE operation='PARTITION RANGE';

10) Clean up your environment by removing the index you created and returning the
constraint to its original state.
SQL> @cleanup_star_schema_lab
SQL> set echo on
SQL>
SQL> set timing off
SQL>
SQL> drop index sales_c_state_bjix;
Index dropped.
SQL>
SQL> alter table customers enable novalidate constraint
n s e
customers_pk;
i ce
b l el
Table altered.
fe r a
SQL>
a n s
SQL> connect / as sysdba
o n -tr
an
Connected.
SQL>
SQL> revoke dba from sh;
h s
a deฺ
e )
s Gui
r ฺ
mi dent
Revoke succeeded.
p
g Stuฺ
SQL> d
@ Database
SQL> exit
b i aOracle t h is 11g Enterprise Edition Release
a
11.1.0.6.0 -(jProduction
Disconnected from
se
i a to
b Partitioning, u
A
With the
andoReal
Oracle Label Security, OLAP, Data Mining
ni Application Testing options

to[oracle@edrsr33p1-orcl Star_Schema_Tuning]$
A n
o s e --------------------------------------------------------------
J set echo on
set timing off

set autotrace off
drop index sales_c_state_bjix;
alter table customers enable novalidate constraint customers_pk;
connect / as sysdba
revoke dba from sh;
exit;

n s e
i ce
b l el
fer a
a n s
n - tr
o
s an
) h a deฺ
r ฺ e s Gui
p ฺ mi dent
d g Stu
b i a@ this
i a (ja use
Ab to
o n io
nt
e A
J o s

Practice 7-1: Using System Statistics
In this practice, you manipulate system statistics and show that they are important for the
optimizer to select the correct execution plans.
1. Connected as the oracle user in a terminal session, execute the
sysstats_setup.sh script located in your
$HOME/solutions/System_Stats directory. This script creates a user

called JFV and some tables used throughout this lab. The script also makes sure
that object statistics are correctly gathered.
[oracle@edrsr33p1-orcl ~]$ cd $HOME/solutions/System_Stats
[oracle@edrsr33p1-orcl System_Stats]$
[oracle@edrsr33p1-orcl System_Stats]$ ./sysstats_setup.sh
n s e
ce
SQL*Plus: Release 11.1.0.6.0 - Production on Mon Mar 31 19:11:42
2008
el i
a b l
fe r
a n s
n - tr
Connected to:
o
an
s
Production
) h a deฺ
r ฺ
and Real Application Testing options s Gui
e
p ฺ mi dent
SQL>
d g Stu
a@ this
SQL> connect / as sysdba;
Connected.
b i
SQL>
(ja cascade;
user jfv u
SQL> dropia se
A b to
o
i dropped.
ton
User
A n
o s e SQL>
J SQL> create user jfv identified by jfv default tablespace users
User created.
SQL>
SQL> grant connect, resource, dba to jfv;
Grant succeeded.
SQL>
SQL>
SQL> connect jfv/jfv
Connected.
SQL>
SQL> drop table t purge;
drop table t purge
*
ERROR at line 1:

Practice 7-1: Using System Statistics (continued)
SQL>
SQL> drop table z purge;
drop table z purge
*
ERROR at line 1:
SQL>
SQL> create table t(c number);
Table created.
SQL>
n s e
SQL> insert into t values (1);
i ce
b l el
1 row created.
fer a
SQL>
a n s
SQL> commit;
n - tr
o
Commit complete.
s an
) h a deฺ
SQL>
e
SQL> insert into t select * fromrฺt;
s Gui
ฺ m i ent
1 row created.
d p
g Stud
SQL> b i a@ this
SQL> /
i a (ja use
A b
created. t
o
o
ni /
2 rows
ntoSQL>
e A
J o s 4 rows created.
SQL> /
8 rows created.
SQL> /
16 rows created.
SQL> /
32 rows created.
SQL> /
64 rows created.
SQL> /

128 rows created.
SQL> /
256 rows created.

SQL> /
512 rows created.
SQL> /
1024 rows created.
SQL>
SQL> commit;
n s e
i ce
Commit complete.
b l el
SQL>
fer a
SQL> insert into t select * from t;
a n s
n -tr
o
an
2048 rows created.
SQL>
h s
a deฺ
SQL> /
e )
s Gui
r ฺ
4096 rows created.
p ฺ mi dent
d g Stu
SQL> /
b i a@ this
i a (ja use
8192 rows created.
A b to
i o
SQL> /
n
t o
n 16384 rows created.
e A
Jos SQL> /
32768 rows created.
SQL> /
65536 rows created.
SQL> /
SQL> /
SQL>
SQL> commit;
Commit complete.

SQL>
SQL> create table z(d number);
Table created.
SQL>
SQL> begin
3 insert into z values (i);
4 end loop;
5 commit;
6 end;
7 /

n s e
i ce
SQL>
b l el
SQL> create unique index iz on z(d);
fer a
Index created.
a n s
o n -tr
an
SQL>
SQL> execute dbms_stats.gather_table_stats('JFV','T',cascade=>true);
h s
a deฺ
e )
s Gui
r ฺ
i ent
SQL>
p ฺ m
d g Stud
SQL> execute dbms_stats.gather_table_stats('JFV','Z',cascade=>true);
a@ this completed.
PL/SQL procedurebisuccessfully
i a (ja use
SQL>
SQL> A
b to
n i o
ntoSQL>
e A SQL> connect / as sysdba;
J o s Connected.
SQL>
System altered.
SQL>
System altered.
SQL>
SQL> execute dbms_stats.delete_system_stats;
SQL>
SQL> execute DBMS_STATS.SET_SYSTEM_STATS (pname => 'cpuspeednw',
pvalue => 0);

SQL>
SQL> select sname,pname,pval1 from aux_stats$;
SNAME PNAME
PVAL1
------------------------------ ------------------------------ ------

----
SYSSTATS_INFO STATUS
SYSSTATS_INFO DSTART
SYSSTATS_INFO DSTOP
SYSSTATS_INFO FLAGS
1
SYSSTATS_MAIN CPUSPEEDNW
0
SYSSTATS_MAIN IOSEEKTIM
n s e
10
i ce
SYSSTATS_MAIN IOTFRSPEED
b l el
4096
SYSSTATS_MAIN SREADTIM
fer a
SYSSTATS_MAIN MREADTIM
a n s
SYSSTATS_MAIN CPUSPEED
o n -tr
an
SYSSTATS_MAIN MBRC
SNAME PNAME
h s
a deฺ
PVAL1
e )
s Gui
i r ฺ n t
------------------------------ ------------------------------ ------
---- m
pฺ MAXTHR de
SYSSTATS_MAIN
d g t u
SYSSTATS_MAIN
i a @ his SSLAVETHR
( j ab se t
13 rows selected.
b ia to u
SQL> A
o n io exit;
SQL>
nt Disconnected from Oracle Database 11g Enterprise Edition Release
e A
J o s 11.1.0.6.0 - Production
--------------------------------------------------------------
#!/bin/bash
cd /home/oracle/solutions/System_Stats
export
sqlplus / as sysdba @sysstats_setup.sql

--------------------------------------------------------------
set echo on
connect / as sysdba;
create user jfv identified by jfv default tablespace users temporary

tablespace temp;
grant connect, resource, dba to jfv;
n s e
connect jfv/jfv
i ce
b l el
drop table t purge;
fer a
drop table z purge;
a n s
n -tr
o
an
create table t(c number);
insert into t values (1);

h s
a deฺ
e )
s Gui
commit; r ฺ
i ent
p ฺ m
d g t;Stud
insert into t select * from
/ b i a@ this
/
i a (ja use
/
A b to
/ni
/
o
t
n /o
e A
Jos
/
/
/
/
commit;
insert into t select * from t;
/
/
/
/
/
/
/
commit;
create table z(d number);

begin
insert into z values (i);
end loop;
commit;
end;
create unique index iz on z(d);
execute dbms_stats.gather_table_stats('JFV','T',cascade=>true);
execute dbms_stats.gather_table_stats('JFV','Z',cascade=>true);
n s e
connect / as sysdba;
i ce
b l el
fe r a
a n s
o n -tr
execute dbms_stats.delete_system_stats;
n
a 'cpuspeednw',
execute DBMS_STATS.SET_SYSTEM_STATS (pname
h a s=>
e ฺ pvalue
=> 0);
s) Gui d
r ฺ
i ente
select sname,pname,pval1 from
p ฺ maux_stats$;
d
d g S t u
exit;
b i a@ this
ja use
2. Frombyouria (terminal
t o session, connect as the JFV user in the SQL*Plus session.
A
ioAfter this, execute the following statement and determine how long it takes to
t o n execute:
A n
o s e select /* Without system stats */ count(*)
from t,z
J where t.c=z.d;
[oracle@edrsr33p1-orcl System_Stats]$ sqlplus jfv/jfv

2008
Connected to:
Production
SQL> @select_without_sysstats
SQL>
SQL> set timing on
SQL>

SQL> select /* Without system stats */ count(*)
2 from t,z
3 where t.c=z.d;
COUNT(*)
----------
524288
Elapsed: 00:00:00.24
SQL>
SQL> set timing off
SQL>
SQL>
--------------------------------------------------------------
set echo on
n s e
ice
set timing on
b l el
select /* Without system stats */ count(*)
fe r a
from t,z
a n s
where t.c=z.d;
o n -tr
set timing off
s an
) h a deฺ
3. Determine the execution plan used to r ฺ e
execute ui statement. In addition,
s theGprevious
determine the optimizer’s cost,ฺm
i
CPU cost,eand ntI/O cost for the previous execution.
What do you observe? dg
p tu d
@use CPU
anot S
is costing. This is because system statistics were
b
a) The optimizer doesi t h
(jathe first
deleted during
i a u e of this lab. The plan chosen by the optimizer might
sstep
notA to
bebthe best one.
o
ni @show_latest_exec_plan
t oSQL>
n SQL> set echo on
e A
Jos
SQL>
SQL> select * from table(dbms_xplan.display_cursor);
PLAN_TABLE_OUTPUT
--------------------------------------------------------------------
------------
SQL_ID 6avdu58tamzju, child number 0
-------------------------------------
select /* Without system stats */ count(*) from t,z where t.c=z.d
------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost |
------------------------------------------------------------
| 0 | SELECT STATEMENT | | | | 134 |
| 1 | SORT AGGREGATE | | 1 | 6 | |
PLAN_TABLE_OUTPUT
--------------------------------------------------------------------
------------

| 2 | NESTED LOOPS | | 524K| 3072K| 134 |
| 3 | TABLE ACCESS FULL| T | 524K| 1536K| 134 |
|* 4 | INDEX UNIQUE SCAN| IZ | 1 | 3 | |
------------------------------------------------------------

---------------------------------------------------
4 - access("T"."C"="Z"."D")
Note
PLAN_TABLE_OUTPUT
--------------------------------------------------------------------
------------
-----
n s e
- cpu costing is off (consider enabling it)
i ce
b l el
25 rows selected.
fer a
a n s
SQL>
n - tr
o
an
SQL> col operations format a20
SQL> col object_name format a11
SQL> col options format a15
h s
a deฺ
SQL> col cost_cpu_io format a30
e )
s Gui
r ฺ
mi dent
SQL>
SQL>
p ฺ
g Stu
d
SQL> select operation operations, object_name, options,
2
b i a@ this
cost||' -- '||cpu_cost||' -- '||io_cost cost_cpu_io
3 from (select * from v$sql_plan where address in (select address
4
i a (ja use from v$sql
5
Ab to where sql_text
o 6io
like '%system stats%' and
n sql_text
nt
e A not like '%connect%'));
J o s OPERATIONS OBJECT_NAME OPTIONS COST_CPU_IO

-------------------- ----------- --------------- -------------------
-----------
SELECT STATEMENT 134 -- --
SORT AGGREGATE -- --
NESTED LOOPS 134 -- -- 134
TABLE ACCESS T FULL 134 -- -- 134
INDEX IZ UNIQUE SCAN -- --
SQL>
--------------------------------------------------------------
set echo on
select * from table(dbms_xplan.display_cursor);
col operations format a20

col object_name format a11
col options format a15

col cost_cpu_io format a30
select operation operations, object_name, options,

from (select * from v$sql_plan where address in (select address
from v$sql
where sql_text like
'%system stats%' and
sql_text not
like '%connect%'));
4. How can you ensure that the optimizer finds a better plan during future executions
of the same statement? Implement your solution.
b) Gather system statistics again. Because you do not have yet a real workload, you n s e
ice
can gather system statistics in NOWORKLOAD mode.
b l el
SQL> connect / as sysdba;
fe r a
Connected.
a n s
SQL> @gather_system_stats
o n -tr
an
SQL> set echo on
SQL>
h s
a deฺ
SQL> execute DBMS_STATS.GATHER_SYSTEM_STATS(gathering_mode =>
'NOWORKLOAD');
e )
s Gui
r ฺ
i ent
ฺ
PL/SQL procedure successfully
p mcompleted.
d
d g S t u
SQL>
b i a@ this from aux_stats$;
SQL> select sname,pname,pval1
SNAME bia
(ja use
PVAL1 A to PNAME
o n io
------------------------------ ------------------------------ ------
nt
e A ----
J o s SYSSTATS_INFO
SYSSTATS_INFO
STATUS
DSTART
SYSSTATS_INFO DSTOP
SYSSTATS_INFO FLAGS
1
SYSSTATS_MAIN CPUSPEEDNW
1893.021
SYSSTATS_MAIN IOSEEKTIM
8.043
SYSSTATS_MAIN IOTFRSPEED
4096
SYSSTATS_MAIN SREADTIM
SYSSTATS_MAIN MREADTIM
SYSSTATS_MAIN CPUSPEED
SYSSTATS_MAIN MBRC
SNAME PNAME
PVAL1
------------------------------ ------------------------------ ------
----
SYSSTATS_MAIN MAXTHR

SYSSTATS_MAIN SLAVETHR
13 rows selected.
SQL>
--------------------------------------------------------------
set echo on
execute DBMS_STATS.GATHER_SYSTEM_STATS(gathering_mode =>

'NOWORKLOAD');
select sname,pname,pval1 from aux_stats$;
n s e
5. Before verifying your solution, you should flush the System Global Area (SGA)
ice
pools, such as the shared pool and the buffer cache. This is done to prevent
b l el
interferences from the previous steps.
fe ra
SQL> @flush_sga
a n s
SQL>
n - tr
o
an
SQL> set echo on
SQL>
h s
a deฺ
e )
s Gui
r ฺ
mi dent
System altered.
p ฺ
g Stu
SQL>
d
b i a@ this
a (ja use
System altered.
i
A b to
ni o
SQL>
ntoSQL>
e A
J o s --------------------------------------------------------------
set echo on
6. Connected as the JFV user again, check your solution against the following
query:
select /* With system stats */ count(*)
from t,z
where t.c=z.d;
c) The optimizer can make a better decision because it was able to use meaningful
system statistics. You can see that the execution time is now half the value it was
previously, and the execution plan now includes CPU costing information.

SQL> connect jfv/jfv
Connected.
SQL> @select_with_sysstats
SQL> set timing on
SQL>
SQL> select /* With system stats */ count(*)
2 from t,z
3 where t.c=z.d;
COUNT(*)
----------
524288
Elapsed: 00:00:00.11
SQL>
SQL> set timing off
n s e
SQL>
i ce
SQL> @show_latest_exec_plan
b l el
SQL> set echo on
fer a
SQL>
a n s
o n -tr
an
PLAN_TABLE_OUTPUT
------------
h s
--------------------------------------------------------------------
a deฺ
SQL_ID 2x55txn3742by, child number 0
e )
s Gui
r ฺ
mi dent
-------------------------------------
p ฺ
select /* With system stats */ count(*) from t,z where t.c=z.d
g Stu
d
b i a@ this
a (ja use
--------------------------------------------------------------------
i
--------
A b to
n i
| Id o | Operation | Name | Rows | Bytes | Cost (%CPU)|
ntoTime |
e A --------------------------------------------------------------------
J o s --------
| 0 | SELECT STATEMENT | | | | 272 (100)|
|
| 1 | SORT AGGREGATE | | 1 | 6 | |
|
PLAN_TABLE_OUTPUT
--------------------------------------------------------------------
------------
|* 2 | HASH JOIN | | 524K| 3072K| 272 (3)|
00:00:03 |
| 3 | INDEX FULL SCAN | IZ | 100 | 300 | 1 (0)|
00:00:01 |
| 4 | TABLE ACCESS FULL| T | 524K| 1536K| 267 (2)|
00:00:03 |
--------------------------------------------------------------------
--------

---------------------------------------------------

2 - access("T"."C"="Z"."D")
21 rows selected.
SQL>
SQL> col operations format a20

SQL> col cost_cpu_io format a30
SQL>
SQL>
SQL> select operation operations, object_name, options,
2 cost||' -- '||cpu_cost||' -- '||io_cost cost_cpu_io
3 from (select * from v$sql_plan where address in (select address
4 from v$sql
n s e
5 where sql_text
i ce
like '%system stats%' and
b l el
6
not like '%connect%'));
sql_text
fer a
a n s
OPERATIONS OBJECT_NAME OPTIONS -tr
COST_CPU_IO
o n
a n 272 -- --
-------------------- ----------- --------------- -------------------
-----------
SELECT STATEMENT
h s
a deฺ-- --
SORT )
s Gui 272 -- 146844065 --
AGGREGATE
e
HASH JOIN r ฺ
i ent
264
IZ g p ฺ m d SCAN
INDEX
d S t u
FULL 1 -- 27121 -- 1
TABLE ACCESS
263 b i a@ this
T FULL 267 -- 84867339 --
i a (ja use
SQL>
Ab to
o n io
--------------------------------------------------------------
nt
e A
J o s set timing on
select /* With system stats */ count(*)

from t,z
where t.c=z.d;
set timing off
--------------------------------------------------------------
set echo on
col operations format a20

col cost_cpu_io format a30
select operation operations, object_name, options,

from (select * from v$sql_plan where address in (select address
from v$sql
where sql_text like
'%system stats%' and
sql_text not
like '%connect%'));
7. Exit from your SQL*Plus session and clean up your environment for this lab by
executing the systats_cleanup.sh script.
SQL> exit
n s e
i ce
b l el
[oracle@edrsr33p1-orcl System_Stats]$ ./sysstats_cleanup.sh
fer a
a n s
2008
o n -tr
n
a reserved.
h s
a deฺ
e )
s Gui
r ฺ
miEdition t
Connected to:
p
Oracle Database 11g Enterprise ฺ d e nRelease 11.1.0.6.0 -
Production g
dOracle S t u
@ s
thi options
ia Testing
a b
j e
SQL> b ia ( to us
i
SQL> A user jfv cascade;
o drop
to n
A n User dropped.
o s e
J SQL>
System altered.
SQL>
SQL> exit;
--------------------------------------------------------------
#!/bin/bash
cd /home/oracle/solutions/System_Stats

export
sqlplus / as sysdba @sysstats_cleanup.sql
--------------------------------------------------------------
set echo on

n s e
i ce
exit;
b l el
fer a
a n s
n - tr
o
s an
) h a deฺ
r ฺ e s Gui
p ฺ mi dent
d g Stu
b i a@ this
i a (ja use
Ab to
o n io
nt
e A
J o s

Practice 7-2: Automatic Statistics Gathering
In this practice, you manipulate object statistics to see their incidences on the execution
plans of your SQL statements. Note: All scripts needed for this lab can be found in your
$HOME/solutions/Automatic_Gather_Stats directory.
1) From a terminal window connected as the oracle user, set up your environment for
this lab by executing the ags_setup.sh script. This script creates a user called
AGS that you use throughout this lab. The script also creates a table called EMP and
an index.
[oracle@edrsr33p1-orcl Automatic_Gather_Stats]$ ./ags_setup.sh
SQL*Plus: Release 11.1.0.6.0 - Production on Mon Apr 7 15:55:28 2008
n s e
ce
el i
a b l
Connected to:
fe r
a n s
Production
n - tr
o
s an
) h a deฺ
SQL>
SQL> drop user ags cascade;
r ฺ e s Gui
drop user ags cascade
p ฺ mi dent
*
d g Stu
a@doesthinot
ERROR at line 1:
b i
ORA-01918: user 'AGS' s exist
i a (ja use
SQL> Ab to
o
i create user ags identified by ags;
ton
SQL>
A n
o s e User created.
J SQL>
SQL> grant dba to ags;
Grant succeeded.
SQL>
System altered.
SQL>
SQL> --
SQL> -- Turn off AUTOTASK
SQL> --
SQL>
SQL> alter system set "_enable_automatic_maintenance"=0;
System altered.

Practice 7-2: Automatic Statistics Gathering (continued)
SQL>
SQL> connect ags/ags
Connected.
SQL>
*
ERROR at line 1:
SQL>
SQL> create table emp
2 (
3 empno number,
4 ename varchar2(20),
n s e
5 phone varchar2(20),
i ce
6 deptno number
b l el
7 );
fer a
Table created.
a n s
n - tr
o
an
SQL>
SQL>
SQL> insert into emp
h s
a deฺ
2 with tdata as
e )
s Gui
r ฺ
mi dent
3 (select rownum empno
4 ฺ
from all_objects
p
g Stu
5
d
where rownum <= 1000)
6
7
select rownum,
b i a@ this
dbms_random.string ('u', 20),
8
i a (ja use
9
Ab case
to
o
10
n
10)io when rownum/100000 <= 0.001 then mod(rownum,
nt
e A 11 else 10
J o s 12
13
end
from tdata a, tdata b
14 where rownum <= 100000;
SQL>
SQL> commit;
Commit complete.
SQL>
SQL> create index emp_i1 on emp(deptno);
Index created.
SQL>
SQL> exec dbms_stats.delete_schema_stats('AGS');

SQL>
SQL> exit;
[oracle@edrsr33p1-orcl Automatic_Gather_Stats]$
--------------------------------------------------------------
#!/bin/bash
cd /home/oracle/solutions/Automatic_Gather_Stats
n s e
i ce
b l el
export
fer
PATH=/u01/app/oracle/product/11.1.0/db_1/bin:/bin:/usr/bin:/usr/locaa
a n s
o n -tr
an
sqlplus / as sysdba @ags_setup.sql
h s
a deฺ
--------------------------------------------------------------
e )
s Gui
set echo on r ฺ
i ent
p ฺ m
drop user ags cascade; g
d S tud
create user ags b i a@ thibys ags;
identified
a
(jags; se
b
grant dba i a
to
to u
o A
ni system flush shared_pool;
toalter
A n
o s e --
J -- Turn off AUTOTASK
--
alter system set "_enable_automatic_maintenance"=0;
connect ags/ags
drop table emp purge;
create table emp

(
empno number,
ename varchar2(20),
phone varchar2(20),
deptno number
);
insert into emp

with tdata as

(select rownum empno
from all_objects
select rownum,
case
when rownum/100000 <= 0.001 then mod(rownum, 10)
else 10
end
where rownum <= 100000;
commit;
create index emp_i1 on emp(deptno);

n s e
i ce
exec dbms_stats.delete_schema_stats('AGS');
b l el
exit;
fe r a
a n s
2) Connect as the AGS user under a SQL*Plus session. o n -tr
s a nsqlplus ags/ags
) h a deฺ
e s GonuMon i Apr 7 15:55:47 2008

r ฺ
p ฺ mi dAll e t
nrights
Copyright (c) 1982, 2007,
d g Stu
Oracle. reserved.
b i a@ this
Connected to:
i a (ja 11guEnterprise
se
b
Oracle Database
A to Edition Release 11.1.0.6.0 -
o n io the Partitioning, Oracle Label Security, OLAP, Data Mining

Production
With
nt and Real Application Testing options
e A
J o s SQL>
3) Determine the distribution of the deptno values found in the EMP table. What do
you observe?
a) You can see that there are 11 different department values, each repeating 0.01%
of the time except value 10 that repeats 99.9% of the time.
SQL> @show_deptno_distribution
SQL>
SQL> select deptno, count(*) cnt_per_deptno, (count(*)*100)/nr
deptno_percent
2 from emp, (select max(empno) nr
3 from emp)
4 group by deptno, nr
5 order by deptno;
DEPTNO CNT_PER_DEPTNO DEPTNO_PERCENT

---------- -------------- --------------
0 10 .01

1 10 .01
2 10 .01
3 10 .01
4 10 .01
5 10 .01
6 10 .01
7 10 .01
8 10 .01
9 10 .01
10 99900 99.9
11 rows selected.
SQL>
--------------------------------------------------------------
n s e
i ce
set echo on
b l el
select deptno, count(*) cnt_per_deptno, (count(*)*100)/nr
fe r a
deptno_percent
a n s
from emp, (select max(empno) nr
o n -tr
an
from emp)
group by deptno, nr
order by deptno;
h s
a deฺ
e )
s Gui
r ฺ
ฺ mi doneany
4) Determine if there are histograms available
p nt column of the EMP table. What do
you observe? d g Stu
i @histograms
ano h is defined on any column of the EMP table.
a) Currently, therebare t
i a (ja use
b echo onto
SQL> @check_emp_histogram
A
o
SQL> set
ni select column_name, histogram, num_buckets
t oSQL>
e An 2 from user_tab_columns
SQL>
Jos 3 where table_name='EMP';
COLUMN_NAME HISTOGRAM NUM_BUCKETS

------------------------------ --------------- -----------
EMPNO NONE
ENAME NONE
PHONE NONE
DEPTNO NONE
SQL>
--------------------------------------------------------------
set echo on
select column_name, histogram, num_buckets

from user_tab_columns
where table_name='EMP';

5) Determine if you currently have statistics gathered on your EMP table. What do you
observe?
a) Currently, there are no statistics gathered on your EMP table.
SQL> @check_table_last_analyzed
SQL> set echo on

SQL>
SQL>
SQL> SELECT
last_analyzed,sample_size,monitoring,num_rows,blocks,table_name
2 FROM user_tables
3 WHERE table_name = 'EMP';
LAST_ANAL SAMPLE_SIZE MON NUM_ROWS BLOCKS TABLE_NAME

n s e
--------- ----------- --- ---------- ---------- --------------------
i ce
----------
b l el
YES EMP
fer a
SQL>
a n s
o n -tr
n
--------------------------------------------------------------
a
set echo on
h s
a deฺ
)
s Gui
set linesize 200 pagesize 1000 rฺe
ฺ m i ent
SELECT
d g p t u d
@ his S
FROM user_tablesbia t
( j a = 'EMP';
s e
ia to u
WHERE table_name
b
n io A if you currently have statistics gathered on the index created on top of the
6)oDetermine
t
e An EMP table. What do you observe?
Jos a) Currently, EMP_I1 has no statistics gathered.
SQL> @check_index_last_analyzed
SQL> set echo on
SQL>
SQL>
SQL> SELECT index_name name,num_rows,
2 last_analyzed,distinct_keys,leaf_blocks,
3 avg_leaf_blocks_per_key,join_index
4 FROM user_indexes
NAME NUM_ROWS LAST_ANAL DISTINCT_KEYS

LEAF_BLOCKS AVG_LEAF_BLOCKS_PER_KEY JOI
------------------------------ ---------- --------- ------------- --
--------- ----------------------- ---
EMP_I1
NO

SQL>
--------------------------------------------------------------
set echo on
SELECT index_name name,num_rows,

last_analyzed,distinct_keys,leaf_blocks,
avg_leaf_blocks_per_key,join_index
FROM user_indexes
WHERE table_name = 'EMP';
7) Execute the following two statements and determine their execution plans:
n s e
select count(*), max(empno) from emp where deptno = 9;
ice
b l el
What do you observe and why?
f a
er
a) You see that for literal 9, the optimizer decided to use the indexra
s
n for literal
whereas
10, the optimizer decided to do TABLE ACCESS FULL. - t
n optimizer
The
n o
a dynamic sampling was
determined the correct plans in both cases. This is because
sobjects.
used, as there were no statistics gathered on h
) a
both
i d eฺ
SQL> @check_exec_plans
i r ฺ es t Gu
SQL> set echo on
g p ฺm uden
SQL>
d S t
SQL> set linesize 200 @ pagesize
s 1000
SQL>
j a b ia
e thi
ia ( to us
SQL>
b
n i o A count(*), max(empno) from emp where deptno = 9;
SQL> select
nto COUNT(*) MAX(EMPNO)

e A
J o s ---------- ----------
10 99
Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
-------------------
(%CPU)| Time |
--------------------------------------------------------------------
-------------------
2 (0)| 00:00:01 |
| 1 | SORT AGGREGATE | | 1 | 26 |
| |
| 2 | TABLE ACCESS BY INDEX ROWID| EMP | 10 | 260 |
2 (0)| 00:00:01 |

|* 3 | INDEX RANGE SCAN | EMP_I1 | 10 | |
1 (0)| 00:00:01 |
--------------------------------------------------------------------
-------------------

---------------------------------------------------
Note
-----
Statistics
n s e
----------------------------------------------------------
i ce
9 recursive calls
b l el
0 db block gets
64 consistent gets
fer a
1 physical reads
a n s
0 redo size
o n -tr
an
h
2 SQL*Net roundtrips to/from client s
a deฺ
0 sorts (memory)
e )
s Gui
r ฺ
mi dent
0 sorts (disk)
1 rows processed
p
g Stuฺ
d
SQL>
b i a@max(empno)
SQL> select count(*), t h is from emp where deptno = 10;
i a (ja use
COUNT(*)
A to
b ----------
MAX(EMPNO)
o
oni 99900 100000
----------
nt
eA
Jos Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
-------
Time |
--------------------------------------------------------------------
-------
00:00:03 |
| 1 | SORT AGGREGATE | | 1 | 26 | |
|
|* 2 | TABLE ACCESS FULL| EMP | 86262 | 2190K| 265 (1)|
00:00:03 |
--------------------------------------------------------------------
-------

---------------------------------------------------
2 - filter("DEPTNO"=10)
Note
-----
Statistics
----------------------------------------------------------
7 recursive calls
0 db block gets
871 consistent gets
5 physical reads
0 redo size
n s e
i ce
b l el
0 sorts (memory)
fer a
0 sorts (disk)
a n s
1 rows processed
o n -tr
SQL>
s an
) h a deฺ
SQL>
r ฺ e s Gui
ฺ mi dent
--------------------------------------------------------------
p
d g Stu
set echo on
b i a@ this
a pagesize
set linesize(j200 se 1000
b i a o u
A ont
o
ni count(*), max(empno) from emp where deptno = 9;
set autotrace
ntoselect
e A
J o s select count(*), max(empno) from emp where deptno = 10;
set autotrace off
8) Confirm your assumption from the previous step.

SQL> show parameter optimizer_dynamic_sampling
NAME TYPE VALUE

------------------------------------ ----------- -------------------
-----------
optimizer_dynamic_sampling integer 2
SQL>
9) Make sure you disable the mechanism found in the previous step.
SQL> alter session set optimizer_dynamic_sampling=0;
Session altered.
SQL>

10) Perform step 7 again. What do you observe and why?
a) Because dynamic sampling is not used, the optimizer cannot use any statistics. It
uses default statistics that are not a good choice for the second statement.
SQL> set echo on

SQL>
SQL>
SQL>
SQL> select count(*), max(empno) from emp where deptno = 9;
COUNT(*) MAX(EMPNO)
---------- ----------
n s e
10 99
i ce
b l el
Execution Plan
fe r a
----------------------------------------------------------
a n s
o n -tr
an
--------------------------------------------------------------------
s
-------------------
|h
a | eRows
Name id
ฺ | Bytes | Cost
| Id | Operation )
es t Gu
(%CPU)| Time |
i r ฺ
g p ฺm uden
--------------------------------------------------------------------
-------------------
d S t |
(0)| 00:00:01ia
@ i s | 1 | 26 |
5
( j b e th
|
aAGGREGATE
| 1 | SORT
i a u s | | 1 | 26 |
|
2A
|
o
b TABLE tACCESS
|
o |
5ni (0)| 00:00:01 |
BY INDEX ROWID| EMP | 714 | 18564 |
nto|* 3 | INDEX RANGE SCAN | EMP_I1 | 286 | |

e A 1 (0)| 00:00:01 |
J o s --------------------------------------------------------------------
-------------------

---------------------------------------------------
Statistics
----------------------------------------------------------
1 recursive calls
0 db block gets
3 consistent gets
0 physical reads
0 redo size
0 sorts (memory)

0 sorts (disk)
1 rows processed
SQL>
COUNT(*) MAX(EMPNO)
---------- ----------
99900 100000
Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
n s e
-------------------
i ce
b l el
(%CPU)| Time |
fer
-------------------------------------------------------------------- a
-------------------
a n s
| 0 | SELECT STATEMENT | | -
1 |
n tr 26 |
o
an
5 (0)| 00:00:01 |
| 1 | SORT AGGREGATE | | 1 | 26 |
| |
h s
a deฺ
| 2 | )
s Gui
TABLE ACCESS BY INDEX ROWID| EMP
e
| 714 | 18564 |
r ฺ
mi dent
5 (0)| 00:00:01 |
|* 3 | INDEX RANGE SCAN
p ฺ
g Stu
| EMP_I1 | 286 | |
1 (0)| 00:00:01 |
d
b
-------------------i a@ this
--------------------------------------------------------------------

b
Predicate
A to
Information
o
ni 3 - access("DEPTNO"=10)
---------------------------------------------------
n t o
e A
Jos Statistics
----------------------------------------------------------
1 recursive calls
0 db block gets
954 consistent gets
190 physical reads
0 redo size
0 sorts (memory)
0 sorts (disk)
1 rows processed
SQL>
SQL>
--------------------------------------------------------------

set echo on
set autotrace on
set autotrace off
11) Reset dynamic sampling as it was at the beginning of this lab.

n s e
ce
eli
Session altered.
a b l
fer
SQL>
a n s
n -tr
12) Set the following wrong statistic values for your EMP table:
o
- Set its number of rows to 10.
s an
- Set its number of blocks to 5.
) h a deฺ
r ฺ e s Gui
mi dent
SQL> @set_wrong_stats
SQL> set echo on
SQL> p ฺ
g Stu
d
a@ this
SQL> exec
b i
dbms_stats.set_table_stats('AGS','EMP',null,null,null,10,5);
u se
b
PL/SQL procedure
A to completed.
o n io
SQL>
nt
e A --------------------------------------------------------------
J o s
set echo on
exec dbms_stats.set_table_stats('AGS','EMP',null,null,null,10,5);
13) Check that you modified correctly the statistics of the EMP table.
SQL> set echo on
SQL>
SQL>
SQL> SELECT
2 FROM user_tables

--------- ----------- --- ---------- ---------- --------------------
----------

07-APR-08 2000 YES 10 5 EMP
SQL>
--------------------------------------------------------------
set echo on
SELECT
FROM user_tables
WHERE table_name = 'EMP';
n s e
ice
l el
a) Because there are statistics defined on the EMP table, the optimizer uses them, and
b
fe
not dynamic sampling. However, because the statistics are incorrect, the ra
n s
generated plans are also incorrect, at least for the second statement.
a
o n -tr
an
SQL> set echo on
s
a deฺ
set linesize 200 pagesize 1000 ) h
SQL>
SQL>
r ฺ e s Gui
mi dent
SQL>
p ฺ
g Stufrom emp where deptno = 9;
SQL>
d
SQL>
b i a@ this
select count(*), max(empno)
i a (ja use
COUNT(*) MAX(EMPNO)
b10
----------
A to 99
----------
i o
nton
e A
J o s Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
-------------------
(%CPU)| Time |
--------------------------------------------------------------------
-------------------
2 (0)| 00:00:01 |
| 1 | SORT AGGREGATE | | 1 | 26 |
| |
2 (0)| 00:00:01 |
1 (0)| 00:00:01 |
--------------------------------------------------------------------
-------------------

---------------------------------------------------
Statistics
----------------------------------------------------------
0 recursive calls
0 db block gets
3 consistent gets
0 physical reads
0 redo size
n s e
0 sorts (memory)
i ce
0 sorts (disk)
b l el
1 rows processed
fer a
SQL>
a n s
o n -tr
COUNT(*) MAX(EMPNO)
s an
---------- ----------
) h a deฺ
99900 100000
r ฺ e s Gui
ฺ m i ent
Execution Plan
d g p t u d
a@ this S
----------------------------------------------------------
i a (ja use
b to
--------------------------------------------------------------------
A
o
niId | Operation
-------------------
nto|(%CPU)| | Name | Rows | Bytes | Cost
e A Time |
J o s --------------------------------------------------------------------
-------------------
2 (0)| 00:00:01 |
| 1 | SORT AGGREGATE | | 1 | 26 |
| |
2 (0)| 00:00:01 |
1 (0)| 00:00:01 |
--------------------------------------------------------------------
-------------------

---------------------------------------------------
Statistics
----------------------------------------------------------

0 recursive calls
0 db block gets
805 consistent gets
0 physical reads
0 redo size

0 sorts (memory)
0 sorts (disk)
1 rows processed
SQL>
SQL>
n s e
--------------------------------------------------------------
i ce
b l el
set echo on
fer a
a n s
o n -tr
a n = 9;
set autotrace on
select count(*), max(empno) from emp where

h s
a deฺ
deptno
select count(*), max(empno) fromrฺempe ) ui = 10;

s whereGdeptno
ฺ m i ent
set autotrace off
d p
g Stud
b i a@ this
ja ugather
15) Make sure you(manually
a se statistics on your EMP table and its corresponding
index. b i to
o A
t o ni @gather_stats_manually
SQL>
e An SQL> set echo on

Jos
SQL>
SQL> exec dbms_stats.gather_table_stats('AGS', 'EMP', cascade =>
true);
SQL>
--------------------------------------------------------------
set echo on
exec dbms_stats.gather_table_stats('AGS', 'EMP', cascade => true);
16) Make sure statistics were gathered on your objects.

SQL> set echo on
SQL>

SQL>
SQL> SELECT
2 FROM user_tables

--------- ----------- --- ---------- ---------- --------------------
----------
07-APR-08 100000 YES 100000 874 EMP
SQL>
SQL> set echo on
SQL>
n s e
SQL>
ice
b l el
2
3
last_analyzed,distinct_keys,leaf_blocks,
fe r a
4 FROM user_indexes
a n s
o n -tr
NAME a n DISTINCT_KEYS
NUM_ROWS LAST_ANAL
LEAF_BLOCKS AVG_LEAF_BLOCKS_PER_KEY JOI as
) h i d eฺ ------------- --
------------------------------ ----------
r ฺ s Gu
e---
---------
--------- -----------------------
EMP_I1 ฺ m i e nt 07-APR-08
100000 11
196
d g p
17 NO
t u d
@ s S
SQL>
j a b ia
e thi
SQL> setb a ( onto us
iecho
SQL> A
o n io select column_name, histogram, num_buckets
SQL>
nt 2 from user_tab_columns
e A
J o s 3 where table_name='EMP';

------------------------------ --------------- -----------
EMPNO NONE 1
ENAME NONE 1
PHONE NONE 1
DEPTNO FREQUENCY 7
SQL>

a) Because statistics were correctly gathered on both objects, the optimizer is able to
use correct execution plans for both statements.
SQL> set echo on
SQL>

SQL>
SQL>
COUNT(*) MAX(EMPNO)
---------- ----------
10 99
Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
-------------------
n s e
i ce
(%CPU)| Time |
b l el
-------------------
fer a
--------------------------------------------------------------------
a n s
8 |
2 (0)| 00:00:01 |
n -tr
o
an
| |
| 2 |
h
TABLE ACCESS BY INDEX ROWID| EMP s
a deฺ
| 18 | 144 |
2 (0)| 00:00:01 |
e )
s Gui
r ฺ
mi dent
1 (0)| 00:00:01 |
p ฺ
g Stu
d
--------------------------------------------------------------------
-------------------
b i a@ this
b to
---------------------------------------------------
A
i o
nton 3 - access("DEPTNO"=9)
e A
J o s Statistics
----------------------------------------------------------
0 recursive calls
0 db block gets
3 consistent gets
0 physical reads
0 redo size
0 sorts (memory)
0 sorts (disk)
1 rows processed
SQL>
COUNT(*) MAX(EMPNO)
---------- ----------
99900 100000

Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
-------
Time |
--------------------------------------------------------------------
-------
00:00:03 |
| 1 | SORT AGGREGATE | | 1 | 8 | |
|
n s e
i ce
00:00:03 |
b l el
--------------------------------------------------------------------
-------
fer a
a n s
o n -tr
an
---------------------------------------------------
h s
a deฺ
e )
s Gui
r ฺ
i ent
Statistics
p ฺ m d
d g t u
----------------------------------------------------------
S
0 db b i a@gets
0 recursive
block
calls
t h is
a sereads
805 (jconsistent gets
i a
b 0 redotosize
0 u
physical
i o A
nton 482 bytes sent via SQL*Net to client
e A 420 bytes received via SQL*Net from client
J o s 2
0
sorts (memory)
0 sorts (disk)
1 rows processed
SQL>
SQL>
18) Ensure you delete all statistics previously generated on your objects.
SQL> @delete_stats
SQL> set echo on
SQL>
SQL> exec dbms_stats.delete_schema_stats('AGS');
SQL>
--------------------------------------------------------------
set echo on

exec dbms_stats.delete_schema_stats('AGS');
19) Verify that you no longer have statistics gathered on both objects.
SQL> set echo on
SQL>
SQL>
SQL> SELECT
2 FROM user_tables
n s e
i ce
--------- ----------- --- ---------- ---------- -------------------- b l el
----------
fer a
YES EMP
a n s
n - tr
SQL> o
s an
SQL> set echo on
) h a deฺ
SQL>
SQL>
r
set linesize 200 pagesize 1000ฺ e s Gui
SQL>
p ฺ mi dent
SQL> g Stu
SELECT index_name name,num_rows,
d
a@ this
3
b i
4 (ja use= 'EMP';
FROM user_indexes
i a
5
b
WHERE
A to
table_name
n i o
n t oNAME
LEAF_BLOCKS
NUM_ROWS LAST_ANAL DISTINCT_KEYS
AVG_LEAF_BLOCKS_PER_KEY JOI
e A ------------------------------ ---------- --------- ------------- --
Jos --------- ----------------------- ---
EMP_I1
NO
SQL>
SQL> set echo on
SQL>
SQL> select column_name, histogram, num_buckets
2 from user_tab_columns
3 where table_name='EMP';

------------------------------ --------------- -----------
EMPNO NONE
ENAME NONE
PHONE NONE
DEPTNO NONE
SQL>

20) How would you determine the list of automated tasks that exist on your database?
a) You can use Enterprise Manager Database Control by navigating to the
Automated Maintenance Tasks page (Home > Server> Automated Maintenance
Tasks). On the Automated Maintenance Tasks page, you can see the three
automated tasks implemented by default on your database.
b) Another possibility is to use the DBA_AUTOTASK_TASK table as shown by the

following statement:
SQL> select task_name from dba_autotask_task;
TASK_NAME
----------------------------------------------------------------
gather_stats_prog
auto_space_advisor_prog
n s e
i ce
el
AUTO_SQL_TUNING_PROG
a b l
SQL>
fe r
21) Exit from your SQL*Plus session. a n s
o n -tr
an
SQL> exit;
h s
a deฺ
)
s Gui
e
r ฺ
mi dent
p ฺ
g ofSthe
22) You now want to observe d
the effects tu automatic statistics-gathering feature of
b i
your database. However, a@you dot h is want to wait until the database automatically
not
opens the next
a ja usewindow. So from your terminal session, execute the
(maintenance
i
b script. toThis script forces the execution of the automatic statistics-
A
run_ags.sh
o
t o ni
gathering task.
e An [oracle@edrsr33p1-orcl Automatic_Gather_Stats]$ ./run_ags.sh

Jos SQL*Plus: Release 11.1.0.6.0 - Production on Mon Apr 7 16:00:03 2008
Connected to:
Production
SQL>
SQL> exec dbms_workload_repository.create_snapshot;
SQL>
SQL> variable window varchar2(20);
SQL>
SQL> begin

2 select upper(to_char(sysdate,'fmday'))||'_WINDOW' into :window
from dual;
3 end;
4 /

SQL>
SQL> print window;
WINDOW
--------------------------------
MONDAY_WINDOW
SQL>
SQL> --
n s e
SQL> -- Open the corresponding maintenance window, but with other
i ce
clients disabled
b l el
SQL> --
SQL>
fer a
SQL> alter system set "_enable_automatic_maintenance"=1
a n s
2 /
n -tr
o
System altered.
s an
) h a deฺ
SQL>
r ฺ e s - Gui
mi :window);
nt
SQL> exec dbms_auto_task_admin.disable(
> ฺ
'auto space advisor', null,
p
g S d e
d tu
b i a@ this
SQL>
i a (ja use
A o
b tuning tadvisor',
SQL> exec dbms_auto_task_admin.disable( -
o
>
n io'sql null, :window);
nt PL/SQL procedure successfully completed.

e A
J o s SQL>
SQL>
SQL> exec dbms_scheduler.open_window(:window, null, true);
SQL>
SQL> --
SQL> -- Close the maintenance window when auto optimizer stats
collection is done
SQL> --
SQL>
SQL>
SQL>
SQL> exec dbms_scheduler.close_window(:window);

SQL>
SQL>
SQL> alter system set "_enable_automatic_maintenance"=0
2 /
System altered.
SQL>
SQL> --
SQL> -- Re-enable the other guys so they look like they are enabled
in EM.
SQL> -- Still they will be disabled because we have set the
underscore.
SQL> --
n s e
SQL>
i ce
SQL> exec dbms_auto_task_admin.enable( -
b l el
> 'auto space advisor', null, :window);
fer a
a n s
o n -tr
an
SQL>
SQL> exec dbms_auto_task_admin.enable( -
> 'sql tuning advisor', null, :window);
h s
a deฺ
e )
s Gui
r ฺ
i ent
p ฺ m d
SQL>
d g S t u
SQL> exit;
i @ Database
aOracle h is 11g Enterprise Edition Release
Disconnected from
a b t
11.1.0.6.0 -(jProduction
i a u seOracle Label Security, OLAP, Data Mining
With the
A to Testing options
b Application
Partitioning,
ni o
and Real
nto[oracle@edrsr33p1-orcl Automatic_Gather_Stats]$
e A
J o s --------------------------------------------------------------
#!/bin/bash
export
sqlplus / as sysdba @run_ags.sql
--------------------------------------------------------------
set echo on
exec dbms_workload_repository.create_snapshot;

variable window varchar2(20);
begin
select upper(to_char(sysdate,'fmday'))||'_WINDOW' into :window from
dual;
end;
/
print window;
--
-- Open the corresponding maintenance window, but with other clients
disabled
--
n s e
alter system set "_enable_automatic_maintenance"=1
i ce
/
b l el
exec dbms_auto_task_admin.disable( -
fer a
'auto space advisor', null, :window);
a n s
o n -tr
an
'sql tuning advisor', null, :window);
h s
a deฺ
e )
s null, ui true);
r ฺ
i ent
exec dbms_scheduler.open_window(:window, G
p ฺ m d
--
d g S t u
is done b i a@ twindow
-- Close the maintenance
h is when auto optimizer stats collection
--
i a (ja use
A b to
o
ni dbms_lock.sleep(120);
ntoexec
e A
J o s exec dbms_scheduler.close_window(:window);

/
--
-- Re-enable the other guys so they look like they are enabled in
EM.
-- Still they will be disabled because we have set the underscore.
--
exec dbms_auto_task_admin.enable( -
'sql tuning advisor', null, :window);
exit;

23) Connect again as the AGS user from a SQL*Plus session.
[oracle@edrsr33p1-orcl Automatic_Gather_Stats]$ sqlplus ags/ags
SQL*Plus: Release 11.1.0.6.0 - Production on Mon Apr 7 16:02:44 2008

Connected to:
Production
SQL>
n s e
i ce
24) View the statistics of your objects. What do you observe and why?
b l el
f e ra
a) The statistics were automatically gathered by the database during the maintenance
window. You can also see this directly from the Automated Maintenance
t r a ns Tasks
o n -
page in Enterprise Manager. The important thing is that the database
n
automatically gathered the right statistics and histograms. Depending on your
environment, you may see different sample sizes.s a
) h a deฺ
r ฺ e s Gui
mi1000dent
SQL>
SQL> set linesize 200 pagesize
p
g Stuฺ
SQL> d
SQL> SELECT
b i a@ this
(ja use
i a
2 FROM user_tables
to = 'EMP';
b table_name
A
3 WHERE
o
t ni
oLAST_ANAL SAMPLE_SIZE MON NUM_ROWS BLOCKS TABLE_NAME
n
eA
--------- ----------- --- ---------- ---------- --------------------
Jos
----------
07-APR-08 100000 YES 100000 874 EMP
SQL>
SQL> set echo on
SQL>
SQL>
3 avg_leaf_blocks_per_key,join_index
4 FROM user_indexes
NAME NUM_ROWS LAST_ANAL DISTINCT_KEYS

LEAF_BLOCKS AVG_LEAF_BLOCKS_PER_KEY JOI
------------------------------ ---------- --------- ------------- --
--------- ----------------------- ---
EMP_I1 100000 07-APR-08 11
196 17 NO

SQL>
SQL> set echo on
SQL>

------------------------------ --------------- -----------
EMPNO NONE 1
ENAME NONE 1
PHONE NONE 1
DEPTNO FREQUENCY 5
n s e
SQL>
ice
25) Perform step 7 again. What do you observe and why? b l el
fe r a
a n s
a) The optimizer can make the right decisions for both statements. This is because of
o n -tr
the statistics that were automatically gathered by the database previously.
s an
SQL> set echo on
) h a deฺ
s Gui
SQL>
r ฺ e
SQL>
p ฺ mi dent
SQL> d g Stu
i
SQL> select count(*),
b a@max(empno)
t h is from emp where deptno = 9;
COUNT(*)i a (ja use
MAX(EMPNO)
A
---------- to
b ----------
i o
ton
10 99
A n
o s e Execution Plan
J ----------------------------------------------------------
--------------------------------------------------------------------
-------------------
(%CPU)| Time |
--------------------------------------------------------------------
-------------------
2 (0)| 00:00:01 |
| |
2 (0)| 00:00:01 |
1 (0)| 00:00:01 |
--------------------------------------------------------------------
-------------------

---------------------------------------------------
Statistics
----------------------------------------------------------
0 recursive calls
0 db block gets
3 consistent gets
0 physical reads
0 redo size
n s e
0 sorts (memory)
i ce
0 sorts (disk)
b l el
1 rows processed
fer a
SQL>
a n s
o n -tr
COUNT(*) MAX(EMPNO)
s an
---------- ----------
) h a deฺ
99900 100000
r ฺ e s Gui
ฺ m i ent
Execution Plan
d g p t u d
a@ this S
----------------------------------------------------------
i a (ja use
b to
--------------------------------------------------------------------
A
o
niId | Operation
-------
nto|Time | Name | Rows | Bytes | Cost (%CPU)|
e A |
J o s --------------------------------------------------------------------
-------
00:00:03 |
| 1 | SORT AGGREGATE | | 1 | 8 | |
|
00:00:03 |
--------------------------------------------------------------------
-------

---------------------------------------------------
Statistics
----------------------------------------------------------
0 recursive calls
0 db block gets

805 consistent gets
0 physical reads
0 redo size
0 sorts (memory)
0 sorts (disk)
1 rows processed
SQL>
SQL>

n s e
ce
SQL> exit;
eli
a b l
fer
a n s
o n -tr
27) From your terminal window, clean up your environment by
a nexecuting the
ags_cleanup.sh script.
h a s eฺ
e s ) u id ./ags_cleanup.sh
ฺ
ir ent G
ฺ m
d gp SProduction
-
tud
on Mon Apr 7 16:03:37 2008
Copyright (c) 1982,

b i a@ h is
2007, Oracle.
t All rights reserved.
i a (ja use
A
Connectedb to: to
n i o
t o
An With the Partitioning, Oracle Label Security, OLAP, Data Mining
Production
e
Jos and Real Application Testing options
SQL>
SQL> drop user ags cascade;
User dropped.
SQL>
SQL> alter system set "_enable_automatic_maintenance"=1;
System altered.
SQL>
SQL> exit;
--------------------------------------------------------------

#!/bin/bash
export
sqlplus / as sysdba @ags_cleanup.sql
--------------------------------------------------------------
n s e
i ce
set echo on
b l el
drop user ags cascade;
fer a
a n s
alter system set "_enable_automatic_maintenance"=1;
n - tr
o
exit;
s an
) h a deฺ
r ฺ e s Gui
p ฺ mi dent
d g Stu
b i a@ this
i a (ja use
Ab to
o n io
nt
e A
J o s

n s e
i ce
b l el
fer a
a n s
n - tr
o
s an
) h a deฺ
r ฺ e s Gui
p ฺ mi dent
d g Stu
b i a@ this
i a (ja use
Ab to
o n io
nt
e A
J o s

Practice 8-1: Understanding Adaptive Cusrsor Sharing
In this practice, you experiment with bind variable peeking and adaptive cursor sharing.
1) Connected to a terminal session as the oracle user, execute the acs_setup.sh
script to set up the environment used for this lab. You can locate this script in your
$HOME/solutions/Adaptive_Cursor_Sharing directory.
[oracle@edrsr33p1-orcl ~]$ cd solutions/Adaptive_Cursor_Sharing

[oracle@edrsr33p1-orcl Adaptive_Cursor_Sharing]$ ./acs_setup.sh

2008
n s e
i ce
Connected to:
Oracle Database 11g Enterprise Edition Release 11.1.0.6.0 - b l el
Production
fe ra
a n s
o n -tr
SQL>
s an
SQL> drop user acs cascade;
) h a deฺ
drop user acs cascade
r ฺ e s Gui
miexistdent
*
ERROR at line 1:
ORA-01918: user 'ACS' does p ฺ
g Stu
not
d
b i a@ this
SQL>
a ja use
(user
i toacs identified
b tablespace
SQL> create by acs default tablespace users
io A
temporary temp;
o n
t User created.
A n
o s e
J SQL>
SQL> grant dba, connect to acs;
Grant succeeded.
SQL>
SQL> connect acs/acs
Connected.
SQL>
*
ERROR at line 1:
SQL>
2 (
3 empno number,

(continued)
6 deptno number
7 );
Table created.
SQL>
SQL>
2 with tdata as
4 from all_objects
5
6
select rownum,
n s e
i ce
el
7 dbms_random.string ('u', 20),
a b l
9 case
fer
10
s
when rownum/100000 <= 0.001 then mod(rownum,
a n
-tr
10)
11 else 10
o n
12
13
end
s an
) h a deฺ
r ฺ e s Gui
p ฺ mi dent
SQL> d g Stu
SQL> create index a
i @ on
emp_i1
h isemp(deptno);
b t
i a (ja use
Index created.
A b to
o
ni exec dbms_stats.gather_table_stats(null, 'EMP', METHOD_OPT =>
SQL>
ntoSQL>
e A 'FOR COLUMNS DEPTNO SIZE 10', CASCADE => TRUE);
SQL>
System altered.
SQL>
SQL> exit;
[oracle@edrsr33p1-orcl Adaptive_Cursor_Sharing]$
--------------------------------------------------------------
#!/bin/bash
cd /home/oracle/solutions/Adaptive_Cursor_Sharing

(continued)
export
sqlplus / as sysdba @acs_setup.sql
--------------------------------------------------------------
set echo on
n s e
ce
drop user acs cascade;
eli
create user acs identified by acs default tablespace users temporary
a b l
tablespace temp;
fer
a n s
grant dba, connect to acs;
n -tr
o
connect acs/acs
s an
) h a deฺ
r ฺ e s Gui
create table emp
( ฺ m i ent
empno number,
d g p t u d
ename
a@ this
varchar2(20), S
phone b i
varchar2(20),
i a (ja use
deptno number
);
A b to
o
ni into emp
ntoinsert
e A with tdata as
J o s (select rownum empno

from all_objects
select rownum,
case
else 10
end
exec dbms_stats.gather_table_stats(null, 'EMP', METHOD_OPT => 'FOR

COLUMNS DEPTNO SIZE 10', CASCADE => TRUE);

exit;

(continued)
2) In your terminal session, connect to the SQL*Plus session as the ACS user. Ensure
that you stay connected to the same SQL*Plus session until the end of this lab. After
you get connected, identify the columns of the EMP table that have histograms.
a) Only the DEPTNO column has a 10 buckets histogram.
[oracle@edrsr33p1-orcl Adaptive_Cursor_Sharing]$ sqlplus acs/acs

2008
n s e
ce
Connected to:
eli
Production
a b l
fe r
a n s
o n -tr
SQL>
s an
SQL>
a deฺ
) h
2
3
r ฺ e s Gui
p ฺ miHISTOGRAM
d e nt
COLUMN_NAME
d g S---------------
tu NUM_BUCKETS
i a @ his NONE
------------------------------ -----------
ab se t
EMPNO
ENAME
a ( j u
NONE
PHONE
b i t o NONE
DEPTNO
i o A HEIGHT BALANCED 10
n
toSQL>
A n
o s e
J --------------------------------------------------------------
set echo on

3) Determine the distribution of all the distinct values found in the DEPTNO column of
the EMP table. What do you find?
a) Values distribution is uniform for all of them (0.01%) except for value 10
(99.9%). This is typical of what is called data skew.
SQL> set echo on
SQL>
deptno_percent

(continued)
3 from emp)
5 order by deptno;
DEPTNO CNT_PER_DEPTNO DEPTNO_PERCENT

---------- -------------- --------------
0 10 .01
1 10 .01
2 10 .01
3 10 .01
4 10 .01
5 10 .01
6
7
10
10
.01
.01
n s e
i ce
el
8 10 .01
9 10 .01
a b l
10 99900 99.9
fe r
a n s
-tr
11 rows selected.
o n
SQL>
s an
h a deฺ
--------------------------------------------------------------
)
r ฺ e s Gui
set echo on
p ฺ mi dent
g Stu
select deptno, count(*)dcnt_per_deptno, (count(*)*100)/nr
@ s
thi nr
iamax(empno)
deptno_percent
j
from emp, (selecta b
( emp)us e
b afrom
ideptno, tonr
i o A
group by
ton
order by deptno;
A n
o s e 4) Before you study the adaptive cursor sharing feature, disable its functionality by
J setting the OPTIMIZER_FEATURES_ENABLE session parameter back to
10.2.0.1. After this is done, ensure that you execute the following command in
your SQL*Plus session: set lines 200 pages 10000. This is used in the lab
to print execution plans correctly.
SQL> alter session set optimizer_features_enable="10.2.0.1";
Session altered.
SQL> set lines 200 pages 10000
5) Determine the execution plan for the following statement:

select /*ACS_L9*/ count(*), max(empno)
from emp
where deptno = 9;
What do you notice and why?
a) The optimizer uses an index range scan because value 9 is very selective.

(continued)
SQL> @select_deptno_literal_9
SQL> set echo on
SQL>
SQL> select /*ACS_L9*/ count(*), max(empno)
2 from emp
3 where deptno = 9;
COUNT(*) MAX(EMPNO)
---------- ----------
10 99
SQL>
SQL> set echo on
SQL> n s e
i ce
el
a b l
PLAN_TABLE_OUTPUT
fer
n s
--------------------------------------------------------------------
a
o n -tr
--------------------------------------------------------------------
----------------------------------------------------------------
SQL_ID 64ngy4j55d1z5, child number 0
s an
-------------------------------------
) h a deฺ
s Gui
select /*ACS_L9*/ count(*), max(empno) from emp where deptno = 9
r ฺ e
p ฺ mi dent
d g Stu
--------------------------------------------------------------------
-------------------
b i a@ this
( j
| Id | Operation a | se | Name | Rows | Bytes | Cost
bia to
(%CPU)| Time u
A
--------------------------------------------------------------------
io
t o n
-------------------
A n | 0 | SELECT STATEMENT | | | |
o s e 2 (100)|
|
|
1 | SORT AGGREGATE | | 1 | 16 |
J | |
2 (0)| 00:00:01 |
1 (0)| 00:00:01 |
--------------------------------------------------------------------
-------------------

---------------------------------------------------
20 rows selected.
SQL>
--------------------------------------------------------------

(continued)
set echo on

from emp
where deptno = 9;
6) Determine the execution plan for the following statement:

from emp
where deptno = 10;
What do you notice and why?
a) The optimizer uses a full table scan because value 10 is not a selective value. n s e
i ce
b l el
SQL> set echo on
fe r a
SQL>
SQL> select /*ACS_L10*/ count(*), max(empno)
a n s
2 from emp
n - tr
o
an
3 where deptno = 10;
COUNT(*) MAX(EMPNO)
h s
a deฺ
---------- ----------
e )
s Gui
r ฺ
mi dent
99900 100000
p ฺ
g Stu
SQL> d
@ his
SQL> set echo onbia
j a e t
SQL> (
ia * from s
utable(dbms_xplan.display_cursor);
b t o
oA
SQL> select
n i
ntoPLAN_TABLE_OUTPUT
e A --------------------------------------------------------------------
J o s --------------------------------------------------------------------
----------------------------------------------------------------
SQL_ID 3232j5gkp2u5h, child number 0
-------------------------------------
select /*ACS_L10*/ count(*), max(empno) from emp where deptno = 10
--------------------------------------------------------------------
-------
Time |
--------------------------------------------------------------------
-------
|
| 1 | SORT AGGREGATE | | 1 | 16 | |
|
00:00:03 |

(continued)
--------------------------------------------------------------------
-------

---------------------------------------------------
19 rows selected.
SQL>
--------------------------------------------------------------
n s e
i ce
el
set echo on
a b l
fe r
from emp
a n s
-tr
where deptno = 10;
o n
n set it to value 9,
7) Define a bind variable called DEPTNO in your SQL*Plus s a session,
and execute the following query, and determine ) itsaexecution
h i d eฺplan:
select /*ACS_1*/ count(*), rmax(empno)
i ฺes nt Gu
from emp
g p ฺm ude
where deptno = :deptno; d S t
What do you notice b
@
iawhy? thi
and
s
( j a e
suses
a) Becauseiathe optimizer u bind peeking the first time you execute a statement
to and because
bbind variable,
i A
oexecution plan with index access isforused.
with a this first execution, value 9 is used, the
to n
e An SQL> variable deptno number;
Jos SQL> exec :deptno := 9
SQL> @select_deptno_bind
SQL> set echo on
SQL>
SQL> select /*ACS_1*/ count(*), max(empno)
2 from emp
3 where deptno = :deptno;
COUNT(*) MAX(EMPNO)
---------- ----------
10 99
SQL>
SQL> set echo on
SQL>

(continued)
PLAN_TABLE_OUTPUT
--------------------------------------------------------------------
--------------------------------------------------------------------
----------------------------------------------------------------
SQL_ID 272gr4hapc9w1, child number 0

-------------------------------------
select /*ACS_1*/ count(*), max(empno) from emp where deptno =
:deptno
--------------------------------------------------------------------
-------------------
n s e
i ce
el
(%CPU)| Time |
--------------------------------------------------------------------
a b l
-------------------
fer
a n s
|
2 (100)|
|
|
1 | SORT AGGREGATE | | n -
1 |tr 16 |
o
|
| 2 |
|
TABLE ACCESS BY INDEX ROWID| EMP
s|an 12 | 192 |
2 (0)| 00:00:01 |
) h a deฺ
r ฺ e s Gui
| EMP_I1 | 12 | |
1 (0)| 00:00:01 |
p ฺ mi dent
--------------------------------------------------------------------
------------------- d g Stu
b i a@(identified
t h is by operation id):
i a (ja use
---------------------------------------------------
A b to
i o
ton
3 - access("DEPTNO"=:DEPTNO)
A n
o s e 20 rows selected.
J
SQL>
--------------------------------------------------------------
set echo on
select /*ACS_1*/ count(*), max(empno)

from emp
where deptno = :deptno;
--------------------------------------------------------------
set echo on

(continued)
8) Determine the execution statistics in terms of child cursors, executions, and buffer
gets for the previously executed statement. What do you observe?
a) In V$SQL, only one child cursor exists, and it has been executed only once (first
time ever in this case). Also, the number of buffer gets is small due to the efficient
access path that was used.

SQL> @show_latest_exec_stats
SQL> set echo on
SQL>
SQL> select child_number, executions, buffer_gets
2 from v$sql
3 where sql_text like 'select /*ACS_1%';
n s e
CHILD_NUMBER EXECUTIONS BUFFER_GETS
ice
------------ ---------- -----------
0 1 3 b l el
fe r a
SQL>
a n s
--------------------------------------------------------------o n -tr
s an
set echo on
) h a deฺ
e
select child_number, executions, ฺbuffer_gets
r s Gui
from v$sql
p ฺ mi dent
where sql_text like 'select
d g Stu /*ACS_1%';
b i a@ this
a 8 again,
9) Perform steps 7(jand s ebut this time using 10 as the bind value for DEPTNO.
What do b i a
you observe u
toand why?
o A
t o ni The execution plan is identical. The index path is used although value 10 is not
a)
e An selective. This is because bind peeking only operates the first time you execute
Jos
your statement. Looking at V$SQL, you can clearly see that there is still only one
child cursor associated with your statement. However, this time, the number of
buffer gets was raised significantly due to the full table scan.
SQL> exec :deptno := 10
SQL> set echo on
SQL>
2 from emp
COUNT(*) MAX(EMPNO)
---------- ----------
99900 100000
SQL>

(continued)
SQL> set echo on
SQL>
PLAN_TABLE_OUTPUT
--------------------------------------------------------------------
--------------------------------------------------------------------
----------------------------------------------------------------
-------------------------------------
:deptno

n s e
i ce
el
--------------------------------------------------------------------
-------------------
a b l
fe r
(%CPU)| Time |
a n s
------------------- n - tr
--------------------------------------------------------------------
o
2 (100)| |
| |
s an | |
| 1 | SORT AGGREGATE |
) h a deฺ
| 1 | 16 |
| |
r ฺ e s Gui
|
2
2 |
(0)| 00:00:01 | p ฺ mi dent
TABLE ACCESS BY INDEX ROWID| EMP | 12 | 192 |
|* 3 | d
INDEX RANGE SCANg Stu | EMP_I1 | 12 | |
1 (0)| 00:00:01 |
b i a@ this
i a (ja use
--------------------------------------------------------------------
-------------------
A b to
o n io
nt ---------------------------------------------------
e A
J o s 3 - access("DEPTNO"=:DEPTNO)
20 rows selected.
SQL>
SQL> set echo on
SQL>
2 from v$sql

------------ ---------- -----------
0 2 957
SQL>
10) Before the next step, flush your shared pool to make sure you wipe out all cursor’s
information.

(continued)
System altered.
SQL>

a) The execution plan is a full table scan because you used value 10 as your first
bind value. There is only one child cursor that is created so far to handle your
statement.

n s e
i ce
b l el
SQL> set echo on
fe r a
SQL>
a n s
SQL>
2
select /*ACS_1*/ count(*), max(empno)
from emp n - tr
o
an
COUNT(*) MAX(EMPNO) h s
a deฺ
---------- ----------
e )
s Gui
r ฺ
99900 100000
p ฺ mi dent
SQL> d g Stu
onbi
a@ this
SQL> set echo a
SQL>
i a (j u se
A b * from
SQL> select to table(dbms_xplan.display_cursor);
ni o
ntoPLAN_TABLE_OUTPUT
e A --------------------------------------------------------------------
J o s --------------------------------------------------------------------
----------------------------------------------------------------
-------------------------------------
:deptno
--------------------------------------------------------------------
-------
Time |
--------------------------------------------------------------------
-------
|
| 1 | SORT AGGREGATE | | 1 | 16 | |
|
00:00:03 |

(continued)
--------------------------------------------------------------------
-------

---------------------------------------------------
2 - filter("DEPTNO"=:DEPTNO)
19 rows selected.
SQL>
SQL>
SQL>
set echo on
n s e
ice
el
2 from v$sql
a b l
fe r
a n s
-tr
------------ ---------- -----------
o n
0 1 872
s an
SQL>
) h a deฺ
esyour bind
12) Perform step 9 again, but this time usei9rฺas G ui What do you observe and
why? p ฺ m dent value.
d g Stu
a) Although value 9 is
i a @ his a full table scan is still used. This is because
very selective,
the second time
j b
a you execute
e t your statement, bind peeking is not done. So you
(
iato use tthe
continue s
u child cursor.
same
b
A :deptno := 9 o
n i
SQL>o exec
to
e An PL/SQL procedure successfully completed.
Jos SQL> @select_deptno_bind
SQL> set echo on
SQL>
2 from emp
COUNT(*) MAX(EMPNO)
---------- ----------
10 99
SQL>
SQL> set echo on
SQL>
PLAN_TABLE_OUTPUT

(continued)
--------------------------------------------------------------------
--------------------------------------------------------------------
----------------------------------------------------------------
-------------------------------------

:deptno
--------------------------------------------------------------------
-------
Time |
--------------------------------------------------------------------
n s e
ice
el
-------
a b l
|
fe r
| 1 | SORT AGGREGATE | | 1 | 16 | |
a n s
|
|* 2 | TABLE ACCESS FULL| EMP | 95000 | 1484K| n -240 (1)| tr
o
00:00:03 |
s an
--------------------------------------------------------------------
-------
) h a deฺ
r ฺ e s Gui
ฺ m i eoperation
Predicate Information (identified by
n t id):
gp Stud
---------------------------------------------------
d
i a@ this
b
i a (ja use
to
19 rows bselected.
A
ni o
ntoSQL>
e A SQL> @show_latest_exec_stats
J o s SQL>
SQL>
set echo on

2 from v$sql

------------ ---------- -----------
0 2 1693
SQL>
13) Before the next step, reset your session to use adaptive cursor sharing, and ensure that
you flush your shared pool again.
SQL> alter session set optimizer_features_enable="11.1.0.6";
Session altered.

(continued)
System altered.
SQL>
14) Perform step 12 again. What do you observe, and why?

a) Because this is the first time you execute the statement, bind peeking is used, and
because value 9 is very selective, the index path is used. Only one child cursor is
used to handle this statement.
n s e
SQL> set echo on
ice
SQL>
b l el
fe ra
2
3
from emp
where deptno = :deptno;
a n s
o n -tr
an
COUNT(*) MAX(EMPNO)
---------- ----------
10 99
h s
a deฺ
e )
s Gui
SQL> r ฺ
i ent
SQL> @show_latest_exec_plan ฺm
SQL> set echo on d p
g Stud
SQL>
i @ his
atable(dbms_xplan.display_cursor);
a
SQL> select * from
j b e t
(
ia to u s
b
oA
PLAN_TABLE_OUTPUT
i
--------------------------------------------------------------------
n
nto--------------------------------------------------------------------
e A ----------------------------------------------------------------
J o s SQL_ID 272gr4hapc9w1, child number 0

-------------------------------------
:deptno
--------------------------------------------------------------------
-------------------
(%CPU)| Time |
--------------------------------------------------------------------
-------------------
2 (100)| |
| 1 | SORT AGGREGATE | | 1 | 16 |
| |
2 (0)| 00:00:01 |
1 (0)| 00:00:01 |

(continued)
--------------------------------------------------------------------
-------------------

---------------------------------------------------
20 rows selected.
SQL>
SQL>
SQL>
set echo on
n s e
i ce
el
2 from v$sql
a b l
fe r
a n s
-tr
------------ ---------- -----------
o n
0 1 54
s an
SQL>
) h a deฺ
r ฺ e s 10 G ui bind value. What do you
observe and why? p ฺ mi dent
15) Perform step 14 again, but this time using value as your
a) Although value 10ais@

dg Sthe tusame index path as in the previous step is
b i t h s
not selective,
icurrently
used. Only one
( j a child cursor
s e is needed to represent your statement.
u
bia:deptnoto:= 10
A
SQL> exec
io procedure successfully completed.
t o n
PL/SQL
e An
Jos
SQL> set echo on
SQL>
2 from emp
COUNT(*) MAX(EMPNO)
---------- ----------
99900 100000
SQL>
SQL> set echo on
SQL>
PLAN_TABLE_OUTPUT
--------------------------------------------------------------------
--------------------------------------------------------------------
----------------------------------------------------------------

(continued)
-------------------------------------
:deptno
--------------------------------------------------------------------
-------------------
(%CPU)| Time |
--------------------------------------------------------------------
-------------------
2 (100)| |
| | | |
n s e
ice
el
| 1 | SORT AGGREGATE | | 1 | 16 |
| |
a b l
fe r
2 (0)| 00:00:01 |
a n s
|* 3 |
1
INDEX RANGE SCAN
(0)| 00:00:01 |
| EMP_I1 |
n
1 |
-
|
tr
o
-------------------
s an
--------------------------------------------------------------------
) h a deฺ
Predicate Information (identified by
r ฺ e s Gui id):
operation
p ฺ mi dent
---------------------------------------------------
d g Stu
b i a@ this
i a (ja use
20 rows selected.
A b to
o
ni @show_latest_exec_stats
SQL>
ntoSQL>
e A SQL> set echo on
J o s SQL>
2 from v$sql

------------ ---------- -----------
0 2 1008
SQL>

a) Because you now use adaptive cursor sharing, the system realizes that you benefit
from another child cursor for handling your statement. This time, a full table
access path is used to better handle your statement.

(continued)
SQL> set echo on
SQL>
2 from emp
COUNT(*) MAX(EMPNO)
---------- ----------
99900 100000
SQL>
SQL> set echo on
SQL>
n s e
i ce
el
a b l
PLAN_TABLE_OUTPUT
fer
a n s
--------------------------------------------------------------------
-tr
--------------------------------------------------------------------
n
----------------------------------------------------------------
o
-------------------------------------
s an
h a deฺ
)
:deptno
r ฺ e s Gui
p
Plan hash value: 2083865914 ฺ mi dent
d g Stu
i a@ this
--------------------------------------------------------------------
b
-------
i a (ja use | Name | Rows | Bytes | Cost (%CPU)|
| Id | Operation
Time
A b| to
ni o
--------------------------------------------------------------------
nto-------
e A | 0 | SELECT STATEMENT | | | | 240 (100)|
J o s |
| 1 | SORT AGGREGATE | | 1 | 16 | |
|
00:00:03 |
--------------------------------------------------------------------
-------

---------------------------------------------------
19 rows selected.
SQL>
SQL> set echo on
SQL>

(continued)
2 from v$sql

------------ ---------- -----------
0 2 1008
1 1 821
SQL>
17) Exit your SQL*Plus session, and execute the acs_cleanup.sh script to clean up
your environment.
SQL> exit;
n s e
i ce
el
a b l
fe r
[oracle@edrsr33p1-orcl Adaptive_Cursor_Sharing]$ ./acs_cleanup.sh
a n s
n -tr
o
an
2008
Copyright (c) 1982, 2007, Oracle. All h

s eฺ
a dreserved.
rights
) i
i r ฺes nt Gu
Connected to:
g p ฺm ude
d
Oracle Database 11g Enterprise
S t
Edition Release 11.1.0.6.0 -
@ s
thi Label Security, OLAP, Data Mining
ia Oracle
Production
j a b
e
ia ( to usTesting options
b
i
SQL>o A
n
toSQL> drop user acs cascade;
A n
o s e User dropped.
J
SQL>
System altered.
SQL>
SQL> exit;
[oracle@edrsr33p1-orcl Adaptive_Cursor_Sharing]$

Practice 8-2: Understanding CURSOR_SHARING
In this practice, you investigate the use of the CURSOR_SHARING initialization
parameter.
1) You can find all the necessary scripts for this lab in your
$HOME/solutions/Cursor_Sharing directory. First, you must set up the

environment for this lab by executing the cs_setup.sh script from a terminal
session connected as the oracle user. This script creates a new user called CS and
the EMP table used throughout this lab.
[oracle@edrsr33p1-orcl ~]$ cd solutions/Cursor_Sharing
[oracle@edrsr33p1-orcl Cursor_Sharing]$
[oracle@edrsr33p1-orcl Cursor_Sharing]$ ./cs_setup.sh
n s e
i ce
2008
b l el
fe ra
a n s
Connected to: o n -tr
an
s
Production
) h a deฺ
r ฺ e s Gui
mi dent
SQL> p ฺ
g Stu
d
a@ this
SQL> drop user cs cascade;
b i
User dropped.
i a (ja use
SQL> A
b to
o
ni createtablespace
n t oSQL>
temporary
user cs identified by cs default tablespace users
temp;
e A
Jos User created.
SQL>
SQL> grant dba, connect to cs;
Grant succeeded.
SQL>
SQL> connect cs/cs
Connected.
SQL>
*
ERROR at line 1:
SQL>

Practice 8-2: Understanding CURSOR_SHARING (continued)
2 (
3 empno number,
6 deptno number
7 );
Table created.
SQL>
SQL>
2 with tdata as
4 from all_objects
5 where rownum <= 1000)
n s e
6 select rownum,
i ce
b l el
8
9
case
fer a
10 n s
when rownum/100000 <= 0.001 then mod(rownum,
a
10)
o n -tr
an
11 else 10
12 end
13 from tdata a, tdata b
h s
a deฺ
e )
s Gui
r ฺ
p ฺ mi dent
d g Stu
SQL>
i @ on
aemp_i1 h isemp(deptno);
SQL> create indexb t
i a (ja use
b
Index created.
A to
n i o
ntoSQL>
e A SQL> execute dbms_stats.gather_table_stats(null, 'EMP', cascade =>
J o s true);
SQL>
System altered.
SQL>
SQL> connect / as sysdba
Connected.
SQL>
SQL> shutdown immediate;
Database closed.
Database dismounted.
ORACLE instance shut down.
SQL>
SQL> startup;
ORACLE instance started.

Total System Global Area 845348864 bytes
Fixed Size 1303188 bytes
Variable Size 541068652 bytes
Database Buffers 297795584 bytes
Redo Buffers 5181440 bytes
Database mounted.
Database opened.
SQL>
SQL> exit;
--------------------------------------------------------------
n s e
i ce
#!/bin/bash
b l el
cd /home/oracle/solutions/Cursor_Sharing
fer a
a n s
o n -tr
an
s
) h a deฺ
export
r ฺ e s Gui
mi dent
ฺ
p
sqlplus / as sysdba @ dg Stu
b i a t h is
@cs_setup.sql
i a (ja use
--------------------------------------------------------------
A b on to
o
ni user cs cascade;
set echo
ntodrop
e A
J o s create user cs identified by cs default tablespace users temporary
tablespace temp;
grant dba, connect to cs;
connect cs/cs
create table emp

(
empno number,
ename varchar2(20),
phone varchar2(20),
deptno number
);
insert into emp

with tdata as

(select rownum empno
from all_objects
select rownum,
case
else 10
end
execute dbms_stats.gather_table_stats(null, 'EMP', cascade => true);

n s e
ice
b l el
connect / as sysdba
fe r a
a n s
shutdown immediate;
o n -tr
startup;
s an
) h a deฺ
exit;
r ฺ e s Gui
ฺ m i ent
2) From the same terminal session,
d p
g connect t uasdthe CS user in the SQL*Plus session, and
stay connected to that a @ until
session i S
sthe end of this lab. For formatting reasons, after
you have connected b i t h
a in thesSQL*Plus session, execute the following command:
( j u e
b ia to
n i oA
[oracle@edrsr33p1-orcl Cursor_Sharing]$ sqlplus cs/cs
t o
e An SQL*Plus: Release 11.1.0.6.0 - Production on Mon Mar 31 14:11:38
Jos
2008
Connected to:
Production

SQL>
3) Check the existence of histograms on the columns of the EMP table, and then
determine the data distribution in the DEPTNO column of the EMP table. What do you
observe?

a) Currently, there are no histograms created on the columns of the EMP table. Also,
it is clear that you have data skew in the DEPTNO column. Value 10 repeats most
of the time (99.9%) whereas all other values only repeat 0.01%.
SQL>


------------------------------ --------------- -----------
EMPNO NONE 1
ENAME NONE 1
PHONE
DEPTNO
NONE
NONE
1
1 n s e
ice
SQL>
b l el
fe ra
SQL> set echo on
a n s
SQL>
n - tr
o
deptno_percent
s an
) h a deฺ
s Gui
3 from emp)
r ฺ e
5 order by deptno;
p ฺ mi dent
d
DEPTNO CNT_PER_DEPTNOg DEPTNO_PERCENT
S tu
i a@ 10th--------------
---------- --------------
b is
1a (
0 ja use10 .01
b i2 t o .01
i o A 3 10 .01
t o n 10 .01
A n 5
4 10
10
.01
.01
o s e 6 10 .01
J 7 10 .01
8 10 .01
9 10 .01
10 99900 99.9
11 rows selected.
SQL>
--------------------------------------------------------------
set echo on

--------------------------------------------------------------
set echo on

select deptno, count(*) cnt_per_deptno, (count(*)*100)/nr
deptno_percent
from emp, (select max(empno) nr
from emp)
group by deptno, nr
order by deptno;
4) Before you continue, ensure that you flush your shared pool.
System altered.
SQL>
n s e
ice
5) How would you force your SQL*Plus session to automatically replace statement b l el
ra
literals with bind variables to make sure the same cursor is used independently of the
fe
literal values?
a n s
SQL> alter session set cursor_sharing = force; o n -tr
s an
Session altered.
) h a deฺ
r ฺ e s Gui
mi dent
SQL>
p ฺ
g execute
6) From the same SQL*Plus d
session, S tuthe following two queries, and then
determine how many b i a@ arethgenerated
cursors is to handle these two statements, and what
(
execution plans j a used.seWhat do you observe
were and why?
b ia count(*),
t o u
i o
select A
select /*CS*/
/*CS*/ count(*),
max(empno)
max(empno)
from
from
emp where deptno = 9;
emp where deptno = 10;
t o n
e An a) Because of the previous step, literal values are replaced with bind variables. The
Jos
FORCE option forces the system to share only one child cursor in this case and
use the exact same execution plan (index range scan).
SQL> set echo on
SQL>
SQL> select /*CS*/ count(*), max(empno) from emp where deptno = 9;
COUNT(*) MAX(EMPNO)
---------- ----------
10 99
SQL>
SQL> set echo on
SQL>
COUNT(*) MAX(EMPNO)
---------- ----------
99900 100000

SQL>
SQL> @show_latest_cursors
SQL> set echo on
SQL>
SQL>
SQL> select sql_text,hash_value
2 from v$sql
3 where sql_text like '%select /*CS%';
SQL_TEXT
HASH_VALUE
--------------------------------------------------------------------
-- ----------
select /*CS*/ count(*), max(empno) from emp where deptno =
n s e
:"SYS_B_0" 3434097775
i ce
b l el
SQL> @show_latest_exec_plans
SQL> set echo on
fer a
SQL>
a n s
n -tr
o
an
SQL> col operation format a16
SQL>
h s
a deฺ
)
s Gui
SQL> select address,hash_value,child_number,
e
r ฺ
mi dent
operation,options,object_name
2 from v$sql_plan
p ฺ
g Stu
d
3 where (address,hash_value) in
4
5 from v$sql b i a@ this
(select address,hash_value
6 (ja use
where sql_text like '%select /*CS%');
i a
A b to
o n io
ADDRESS
OBJEC
HASH_VALUE CHILD_NUMBER OPERATION OPTIONS
nt
e A -------- ---------- ------------ ---------------- --------------- --
J o s ---
4C9C53D4 3434097775 0 SELECT STATEMENT
4C9C53D4 3434097775 0 SORT AGGREGATE
4C9C53D4 3434097775 0 TABLE ACCESS BY INDEX ROWID
EMP
4C9C53D4 3434097775 0 INDEX RANGE SCAN
EMP_I
1
SQL>
--------------------------------------------------------------
set echo on
select /*CS*/ count(*), max(empno) from emp where deptno = 9;
--------------------------------------------------------------
set echo on

--------------------------------------------------------------
set echo on
select sql_text,hash_value
from v$sql
where sql_text like '%select /*CS%';
--------------------------------------------------------------
set echo on
n s e
i ce
b l el
col operation format a16
fe r a
a n s
select address,hash_value,child_number,
o n -tr
an
from v$sql_plan
where (address,hash_value) in
h s
a deฺ
e )
s Gui
r ฺ
mi dent
from v$sql
p
g Stuฺ
d
b i
7) Ensure that you create aa @ t h ishistogram on the DEPTNO column of the EMP
10 buckets
table.
i a (ja use
A to
b dbms_stats.gather_table_stats(null,
i o
SQL> exec
n
'EMP', METHOD_OPT =>
nto
'FOR COLUMNS DEPTNO SIZE 10', CASCADE => TRUE);
e A
Jos
SQL> set echo on
SQL>

------------------------------ --------------- -----------
EMPNO NONE 1
ENAME NONE 1
PHONE NONE 1
DEPTNO HEIGHT BALANCED 10
SQL>
--------------------------------------------------------------
set echo on

8) Before you continue, ensure that you flush your shared pool.
System altered.
SQL>
9) Perform step 6 again. What do you observe and why? e

n s
a) Although you captured histogram for the DEPTNO column that shows data skew, i ce
l
the system continues to share only one child cursor to handle both statements.
b el
This behavior is due to the FORCE option for the CURSOR_SHARING fe r a
initialization parameter. a n s
o n -tr
SQL> @select_deptno_literal_9 a n
SQL> set echo on
h s
a deฺ
SQL>
e )
s from i
uemp
i r
SQL> select /*CS*/ count(*), max(empno)ฺ n t G where deptno = 9;
m
COUNT(*) MAX(EMPNO)
d gpฺ Stude
---------- ----------
i @ his
a99
10
j a b e t
(
ia to u s
SQL>
A b
i
SQL>
n o @select_deptno_literal_10
ntoSQL> set echo on
e A SQL>
J o s SQL> select /*CS*/ count(*), max(empno) from emp where deptno = 10;
COUNT(*) MAX(EMPNO)
---------- ----------
99900 100000
SQL>
SQL> set echo on
SQL>
SQL>
2 from v$sql
SQL_TEXT
HASH_VALUE
--------------------------------------------------------------------
-- ----------

:"SYS_B_0" 3434097775
SQL>
--------------------------------------------------------------
set echo on
--------------------------------------------------------------
set echo on

n s e
i ce
--------------------------------------------------------------
b l el
set echo on
fer a
a n s
o n -tr
select sql_text,hash_value
s an
where sql_text like '%select /*CS%'; ) h
from v$sql a deฺ
r ฺ e s Gui
ฺ m i ent
10) Before you continue, ensure that
d p
g you S tudyour shared pool.
flush
SQL> alter systemia @ shared_pool;
h is
b flush
t
i a (ja use
b
System altered.
A to
ni
SQL>o
t o
e An11) How would you ensure that you now use more than one child cursor to handle both
Jos statements? Implement your solution, and check it.
a) By setting CURSOR_SHARING to SIMILAR for your session, the system is able
to see that you benefit from using two different child cursors to handle both
statements because they lend themselves to very different execution plans.
SQL> alter session set cursor_sharing = similar;
Session altered.
SQL> set echo on
SQL>
COUNT(*) MAX(EMPNO)
---------- ----------
10 99

SQL>
SQL> set echo on
SQL>
COUNT(*) MAX(EMPNO)
---------- ----------
99900 100000
SQL>
SQL> set echo on
SQL>
SQL>
n s e
i ce
2 from v$sql
b l el
fer a
SQL_TEXT
a n s
HASH_VALUE
n -tr
o
an
--------------------------------------------------------------------
-- ----------
h s
a deฺ
:"SYS_B_0" 3434097775
e )
s Gui
r ฺ
mi dent
:"SYS_B_0" 3434097775
p ฺ
g Stu
d
@ his
SQL> @show_latest_exec_plans
SQL> set echo onbia t
SQL> ( j a s e
SQL> colb a to u format a5
iobject_name
i A operation format a16
o col
SQL> col
n
toSQL> options format a15
SQL>
A n
o s e SQL> select address,hash_value,child_number,
J operation,options,object_name
2 from v$sql_plan
3 where (address,hash_value) in
4 (select address,hash_value
5 from v$sql
6 where sql_text like '%select /*CS%');
ADDRESS HASH_VALUE CHILD_NUMBER OPERATION OPTIONS

OBJEC
-------- ---------- ------------ ---------------- --------------- --
---
4C9C53D4 3434097775 1 TABLE ACCESS FULL
EMP
4C9C53D4 3434097775 0 TABLE ACCESS BY INDEX ROWID
EMP

4C9C53D4 3434097775 0 INDEX RANGE SCAN
EMP_I
1
7 rows selected.
SQL>
--------------------------------------------------------------
set echo on
--------------------------------------------------------------
n s e
ice
set echo on
b l el
fe r a
a n s
o n -tr
--------------------------------------------------------------
set echo on
s an
) h a deฺ
r ฺ e s Gui
select sql_text,hash_value ฺmi e nt
from v$sql p
dg /*CS%'; tu d
where sql_text like @
a '%select
is S
b i t h
i a (ja use
--------------------------------------------------------------
A b on to
o
ni object_name format a5
set echo
ntocol
e A col operation format a16
J o s col options format a15

select address,hash_value,child_number,
from v$sql_plan
where (address,hash_value) in
from v$sql
12) Exit your SQL*Plus session and execute the cs_cleanup.sh script to clean up
your environment for this lab.
SQL> exit
[oracle@edrsr33p1-orcl Cursor_Sharing]$ ./cs_cleanup.sh

2008

Connected to:
Production
SQL>
SQL> drop user cs cascade;
User dropped.
n s e
i ce
SQL>
b l el
fer a
System altered.
a n s
o n -tr
an
SQL>
SQL> exit;
h s
a deฺ
e )
s Gui
r ฺ
mi dent
p ฺ
g Stu
d
b i a@ this
--------------------------------------------------------------
i a (ja use
b
#!/bin/bash
A to
o
cdi /home/oracle/solutions/Cursor_Sharing
nton
e A export ORACLE_SID=orcl
J o s
export
sqlplus / as sysdba @cs_cleanup.sql
--------------------------------------------------------------
set echo on
drop user cs cascade;
exit;

n s e
i ce
b l el
fer a
a n s
n - tr
o
s an
) h a deฺ
r ฺ e s Gui
p ฺ mi dent
d g Stu
b i a@ this
i a (ja use
Ab to
o n io
nt
e A
J o s

Practice 9-1: Using Hints
In this practice, you study five different hint cases. They are all independent from each
other. Note: You can find all the necessary scripts used for this lab in your
$HOME/solutions/Hints directory. You should be using a terminal session
connected as the oracle user.
1) From your terminal session, execute the iot_setup.sh script. This script creates
an index-organized table.
[oracle@edrsr33p1-orcl Hints]$ ./iot_setup.sh
SQL*Plus: Release 11.1.0.6.0 - Production on Tue Apr 1 15:17:24 2008

n s e
i ce
Connected to:
b l el
fe ra
Production
a n s
-tr
o n
SQL>
s an
SQL> drop user iot cascade;
) h a deฺ
drop user iot cascade
r ฺ e s Gui
*
ERROR at line 1: p ฺ mi dent
d
ORA-01918: user 'IOT' doesg notSexist
tu
b i a@ this
SQL>
i a (ja use
A b usertoiot identified by iot default tablespace users
SQL> create
n i o
nt o
e A User created.
J o s SQL>
SQL> grant connect, resource, dba to iot;
Grant succeeded.
SQL>
SQL> connect iot/iot
Connected.
SQL>
SQL> drop table iottab purge;
drop table iottab purge
*
ERROR at line 1:
SQL>
SQL> CREATE TABLE IOTTAB (
2 OBJECT_ID NUMBER(14, 0) NOT NULL ENABLE
3 , OBJECT_ID_ATT NUMBER(14, 0) NOT NULL ENABLE

Practice 9-1: Using Hints (continued)
4 , OBJECT_ID_CAT NUMBER(14, 0) NOT NULL ENABLE
5 , BEGIN DATE NOT NULL ENABLE
6 , END DATE NOT NULL ENABLE
7 , STATUS NUMBER
8 , COMM VARCHAR2(32) NOT NULL ENABLE
9 , CONSTRAINT IOTTAB_PK
10 PRIMARY KEY (OBJECT_ID

11 , OBJECT_ID_ATT
12 , OBJECT_ID_CAT
13 , BEGIN
14 , END) ENABLE )
15 ORGANIZATION INDEX PCTTHRESHOLD 50 ;
Table created.
SQL>
n s e
SQL> CREATE INDEX OBJECT_ID_ATT_INDX ON IOTTAB (OBJECT_ID_ATT);
i ce
b l el
Index created.
fer a
SQL>
a n s
SQL> -- load data
n -tr
o
an
SQL>
SQL> begin
2 for i in 400001..500000 loop
h s
a deฺ
3 )
s Gui
insert into iottab values(i,mod(i,428),mod(i,20),sysdate-
e
r ฺ
mi dent
mod(i,100),sysdate+mod(i,100),mod(i,3),'aaaaaaaaaaaaaaaaaaaaaaaaaa'|
|i);
p ฺ
g Stu
4 end loop;
d
5 commit;
6 end; b i a@ this
7 /
i a (ja use
A b to
o n io
nt SQL>
e A
J o s SQL>
SQL> begin
2 for i in 100001..200000 loop
3 insert into iottab values(i,mod(i,428),mod(i,20),sysdate-
|i);
4 end loop;
5 commit;
6 end;
7 /
SQL>
SQL>
SQL> begin
2 for i in 300001..400000 loop
|i);
4 end loop;

5 commit;
6 end;
7 /

SQL>
SQL>
SQL> begin
2 for i in 500001..600000 loop
|i);
4 end loop;
5 commit;
6 end;
n s e
7 /
i ce
b l el
fer a
SQL>
a n s
SQL>
n -tr
o
an
SQL> begin
3
h s
a deฺ
)
s Gui
e
r ฺ
mi dent
|i);
4 end loop;
p ฺ
g Stu
5 commit;
d
6 end;
7 / b i a@ this
u se
b
PL/SQL procedure
A to completed.
o n io
SQL>
Ant SQL>
e
Jos
SQL>
System altered.
SQL>
System altered.
SQL>
SQL> exit;
[oracle@edrsr33p1-orcl Hints]$
--------------------------------------------------------------
#!/bin/bash

cd /home/oracle/solutions/Hints
export
sqlplus / as sysdba @iot_setup.sql
--------------------------------------------------------------
set echo on
n s e
i ce
drop user iot cascade;
b l el
fer
create user iot identified by iot default tablespace users temporary a
tablespace temp;
a n s
n -tr
o
an
grant connect, resource, dba to iot;
connect iot/iot
h s
a deฺ
e )
s Gui
drop table iottab purge; r ฺ
i ent
p ฺ m d
CREATE TABLE IOTTAB (
d g S t u
OBJECT_ID
, OBJECT_ID_ATTbia
@ his0)
NUMBER(14, NOT NULL ENABLE
t
NUMBER(14, 0) NOT NULL ENABLE
i a (ja NUMBER(14,
, OBJECT_ID_CAT
u seNOT NULL0) NOT NULL ENABLE
, BEGIN
A b toDATE NOT NULL ENABLE
DATE
n o
,i STATUS
, END ENABLE
nto , COMM NUMBER
e A VARCHAR2(32) NOT NULL ENABLE
J o s , CONSTRAINT IOTTAB_PK
PRIMARY KEY (OBJECT_ID
, OBJECT_ID_ATT
, OBJECT_ID_CAT
, BEGIN
, END) ENABLE )
ORGANIZATION INDEX PCTTHRESHOLD 50 ;
CREATE INDEX OBJECT_ID_ATT_INDX ON IOTTAB (OBJECT_ID_ATT);
-- load data
begin
for i in 400001..500000 loop
|i);
end loop;
commit;
end;
/

begin
for i in 100001..200000 loop
|i);
end loop;
commit;
end;
/
begin
for i in 300001..400000 loop
n s e
i ce
|i);
b l el
end loop;
commit;
fer a
end;
a n s
/
o n -tr
s an
begin
) h a deฺ
for i in 500001..600000 loop
r ฺ e s Gui
ฺ m i ent
|i);
d g p t u d
end loop;
a@ this S
commit; b i
end;
i a (ja use
/
A b to
ni o
ntobegin
e A
J o s for i in 1..100000 loop
|i);
end loop;
commit;
end;
/
exit;

2) From your terminal session, connect as the IOT user in the SQL*Plus session. From
your SQL*Plus session, execute the set_session.sql script. This script sets a
number of parameters for the duration of this case. Ensure that you do not exit from
your SQL*Plus session until asked to do so.
[oracle@edrsr33p1-orcl Hints]$ sqlplus iot/iot
Connected to:
Production
n s e
ce
el i
a b l
SQL> @set_session
fe r
a n s
Session altered.
o n -tr
SQL>
s an
) h a deฺ
r ฺ e s Gui
--------------------------------------------------------------
ฺ mi dent
ALTER SESSION SET OPTIMIZER_INDEX_CACHING=2;
p
d g Stu
a@ th1000
set timing on
set linesize 200 ipagesize
b is
i a (ja use
A
3) From your tosession, execute the following query and note down the time it
b SQL*Plus
n o
i to execute:
takes
t o
e An SELECT comm.
Jos
FROM iottab
WHERE object_id = 1
AND object_id_cat = 0
AND object_id_att = 426 ;
SQL> @select_iot
SQL>
SQL> SELECT comm
2 FROM iottab
3 WHERE object_id = 1
4 AND object_id_cat = 0
5 AND object_id_att = 426 ;
no rows selected
Elapsed: 00:00:00.07
SQL>
--------------------------------------------------------------
set echo on

SELECT comm
FROM iottab
WHERE object_id = 1
4) Use the DBMS_XPLAN package to display the execution plan associated with the
statement you executed in the previous step. What do you observe?
a) It is strange to see that the optimizer chooses a plan that accesses the secondary
index while the query references all columns of the primary key.
SQL> set echo on
n s e
SQL>
i ce
SQL> select * from
b l el
table(dbms_xplan.display_cursor(null,null,'TYPICAL'));
fe r a
PLAN_TABLE_OUTPUT
a n s
o n -tr
--------------------------------------------------------------------
an
--------------------------------------------------------------------
----------------------------------------------------------------
SQL_ID 6u4tsfpprgvzn, child number 0
h s
a deฺ
-------------------------------------
e )
s Gui
r ฺ
mi dent
SELECT comm FROM iottab WHERE object_id = 1 AND object_id_cat
= 0
p ฺ
g Stu
d
AND object_id_att = 426
a@ this
i a (ja use
b to
--------------------------------------------------------------------
A
i o
--------------------
nto|n Id | Operation | Name | Rows | Bytes |
e A Cost (%CPU)| Time |
J o s --------------------------------------------------------------------
--------------------
1 (100)| |
|* 1 | INDEX UNIQUE SCAN| IOTTAB_PK | 1 | 57 |
1 (0)| 00:00:01 |
|* 2 | INDEX RANGE SCAN| OBJECT_ID_ATT_INDX | 50 | |
1 (0)| 00:00:01 |
--------------------------------------------------------------------
--------------------

---------------------------------------------------
1 - access("OBJECT_ID_ATT"=426)
filter(("OBJECT_ID"=1 AND "OBJECT_ID_CAT"=0))
2 - access("OBJECT_ID_ATT"=426)
Note
-----

26 rows selected.
Elapsed: 00:00:00.13
SQL>
--------------------------------------------------------------
set echo on
select * from table(dbms_xplan.display_cursor(null,null,'TYPICAL'));
5) Before trying to fix this issue, make sure you flush the important content of your SGA
by executing the flush_sga.sql script.
n s e
i ce
SQL> @flush_sga
b l el
SQL> set echo on
SQL>
fe r a
a n s
o n -tr
an
System altered.
Elapsed: 00:00:00.11
h s
a deฺ
SQL>
e )
s Gui
r ฺ
i ent
p ฺ m
System altered.
d g Stud
b i
SQL>
i a (ja use
A b to
i o
--------------------------------------------------------------
nton
e A set echo on
J o s alter system flush shared_pool;
6) Using hints only, how would you fix the issue raised at step 4? Implement your
solution and check if it works.
a) The idea is to use a hint to prevent the optimizer from using the secondary index.
You will not use the NO_INDEX hint.
SQL> @select_iot_hint
SQL> set echo on
SQL>
SQL> SELECT /*+ NO_INDEX(t OBJECT_ID_ATT_INDX) */ comm
2 FROM iottab t
3 WHERE object_id = 1
4 AND object_id_cat = 0
5 AND object_id_att = 426 ;

no rows selected
Elapsed: 00:00:00.03
SQL>
SQL> set echo on
SQL>
SQL> select * from
PLAN_TABLE_OUTPUT
--------------------------------------------------------------------
--------------------------------------------------------------------
----------------------------------------------------------------
SQL_ID 4zcxy7z1cg8da, child number 0
-------------------------------------
n s e
SELECT /*+ NO_INDEX(t OBJECT_ID_ATT_INDX) */ comm FROM iottab t
i ce
WHERE object_id = 1 AND object_id_cat = 0 AND object_id_att =
b l el
426
fer a
a n s
n - tr
o
an
--------------------------------------------------------------------
----------
h s
a deฺ
| Rows | Bytes | Cost (%CPU)|
Time |
e )
s Gui
r ฺ
mi dent
--------------------------------------------------------------------
----------
p ฺ
g Stu
|
d
0 | SELECT STATEMENT | | | | 1 (100)|
|
b i a@ this
|* 1 | INDEX RANGE SCAN| IOTTAB_PK | 1 | 57 | 1 (0)|
00:00:01 |
i a (ja use
b to
--------------------------------------------------------------------
n i oA
----------
ntoPredicate Information (identified by operation id):

e A
J o s ---------------------------------------------------
1 - access("OBJECT_ID"=1 AND "OBJECT_ID_ATT"=426 AND

"OBJECT_ID_CAT"=0)
Note
-----
24 rows selected.
Elapsed: 00:00:00.13
SQL>
--------------------------------------------------------------
set echo on
SELECT /*+ NO_INDEX(t OBJECT_ID_ATT_INDX) */ comm

FROM iottab t

WHERE object_id = 1
--------------------------------------------------------------
set echo on
7) Exit from your SQL*Plus session, and clean up your environment by executing the
iot_cleanup.sh script.
SQL> exit;
n s e
i ce
b l el
and Real Application Testing options fe r a
a n s
[oracle@edrsr33p1-orcl Hints]$ ./iot_cleanup.sh
o n -tr
n
SQL*Plus: Release 11.1.0.6.0 - Production onaTue Apr 1 15:21:09 2008
s eฺ
a dreserved.
hrights
Copyright (c) 1982, 2007, Oracle. s ) i
ฺe nt Gu
All
i r
g p ฺm ude
Connected to:
Oracle Database 11g @
d S t
b ia hi s
Enterprise
t
Edition Release 11.1.0.6.0 -
Production
j a e
and Realb a ( to usTesting
iApplication Oracle Label Security, OLAP, Data Mining
options
io A
t o n
SQL>
A n SQL> drop user iot cascade;
o s e
J User dropped.
SQL>
SQL> exit;
--------------------------------------------------------------
#!/bin/bash

export
sqlplus / as sysdba @iot_cleanup.sql

--------------------------------------------------------------
set echo on
drop user iot cascade;
exit;
8) You study a second case of hint utilization to specify hints in lower query blocks.
n s e
From your terminal session connected as the oracle user, execute the
i ce
hr_hint_setup.sh script.
b l el
fe r a
[oracle@edrsr33p1-orcl Hints]$ ./hr_hint_setup.sh
a n s
n -tr
o
n
a reserved.
h s
a deฺ
e )
s Gui
r ฺ
Connected to:
p ฺ miEdition
d e t
nRelease 11.1.0.6.0 -
Production d g Stu
b i
a@Testing isoptions
Oracle
t h Label Security, OLAP, Data Mining
i a (ja use
SQL>
A b o
thr
n i
SQL> o alter user identified by hr account unlock;
nt o
e A User altered.
J o s SQL>
SQL> grant dba to hr
2
SQL> exit;
--------------------------------------------------------------
#!/bin/bash

export
sqlplus / as sysdba @hr_hint_setup.sql

--------------------------------------------------------------
set echo on
alter user hr identified by hr account unlock;
grant dba to hr
exit;
n s e
i ce
9) From your terminal session, connect as the HR user in the SQL*Plus session. After
b l el
fe r a
you are connected, execute the create_hr_view1.sql script that creates a view
a n s
called V1 on top of the EMPLOYEES table. After this is done, execute the
o n -tr
create_hr_view2.sql script that creates a view V2 on top of V1.
[oracle@edrsr33p1-orcl Hints]$ sqlplus hr/hr
s an
h a on Tued ฺ 1 16:09:02 2008
eApr
SQL*Plus: Release 11.1.0.6.0 - Production ) i
i r ฺesAllnrights
t Gu reserved.
g p ฺm ude
Copyright (c) 1982, 2007, Oracle.
d S t
@ s
Connected to:
j a b ia
e thi Edition Release 11.1.0.6.0 -
a ( to us
b
Productioni
i o A Partitioning,
With the Oracle Label Security, OLAP, Data Mining
ton
A n
o s e SQL> @create_hr_view1
J SQL>
SQL> CREATE OR REPLACE VIEW v1 AS
2 SELECT *
3 FROM employees
4 WHERE employee_id < 150;
View created.
SQL>
SQL> @create_hr_view2
SQL> set echo on
SQL>
SQL> CREATE OR REPLACE VIEW v2 AS
2 SELECT v1.employee_id employee_id, departments.department_id
department_id
3 FROM v1, departments
4 WHERE v1.department_id = departments.department_id;
View created.

SQL>
--------------------------------------------------------------
set echo on
CREATE OR REPLACE VIEW v1 AS

SELECT *
FROM employees
WHERE employee_id < 150;
--------------------------------------------------------------
set echo on
CREATE OR REPLACE VIEW v2 AS

n s e
SELECT v1.employee_id employee_id, departments.department_id
i ce
department_id
b l el
FROM v1, departments
WHERE v1.department_id = departments.department_id;
fer a
a n s
n - tr
o
10) Determine the execution plan used to process the following query:
SELECT * FROM v2 WHERE department_id = 30; a n
h a s eฺ
e s ) u id
SQL> @show_exec_plan_view2
ฺ
ir ent G
SQL> set echo on
ฺ m
SQL>
SQL> explain plan for d
gp Stud
2 SELECT *
b i a@ this
3 FROM
a ja use = 30;
(department_id
v2
b
4 WHERE i to
i o A
ton
Explained.
A n
o s e SQL>
J SQL> select * from table(DBMS_XPLAN.DISPLAY(null,null,'ALL'));
PLAN_TABLE_OUTPUT
--------------------------------------------------------------------
--------------------------------------------------------------------
----------------------------------------------------------------
--------------------------------------------------------------------
------------------------------
--------------------------------------------------------------------
------------------------------
33 | 1 (0)| 00:00:01 |
| 1 | NESTED LOOPS | | 3 |
33 | 1 (0)| 00:00:01 |
|* 2 | INDEX UNIQUE SCAN | DEPT_ID_PK | 1 |
4 | 0 (0)| 00:00:01 |

|* 3 | TABLE ACCESS BY INDEX ROWID| EMPLOYEES | 3 |
21 | 1 (0)| 00:00:01 |
|* 4 | INDEX RANGE SCAN | EMP_DEPARTMENT_IX | 6 |
| 0 (0)| 00:00:01 |
--------------------------------------------------------------------
------------------------------

-------------------------------------------------------------
1 - SEL$5C160134
2 - SEL$5C160134 / DEPARTMENTS@SEL$2
3 - SEL$5C160134 / EMPLOYEES@SEL$3
4 - SEL$5C160134 / EMPLOYEES@SEL$3

n s e
---------------------------------------------------
i ce
b l el
2 - access("DEPARTMENTS"."DEPARTMENT_ID"=30)
3 - filter("EMPLOYEE_ID"<150)
fer a
4 - access("DEPARTMENT_ID"=30)
a n s
o n -tr
an
-----------------------------------------------------------
h s
a deฺ
)
s Gui
1 - (#keys=0) "DEPARTMENTS"."DEPARTMENT_ID"[NUMBER,22],
e
"EMPLOYEE_ID"[NUMBER,22] r ฺ
i ent
ฺ m
2 - "DEPARTMENTS"."DEPARTMENT_ID"[NUMBER,22]
p d
d g
3 - "EMPLOYEE_ID"[NUMBER,22]
S t u
b i a@ this
4 - "EMPLOYEES".ROWID[ROWID,10]
i a (ja use
34 rows selected.
SQL> A
b to
ni o
nto--------------------------------------------------------------
e A
J o s set echo on
explain plan for

SELECT *
FROM v2
WHERE department_id = 30;
select * from table(DBMS_XPLAN.DISPLAY(null,null,'ALL'));
11) How do you force the query from step 10 to do a full table scan of the
departments table and a range scan of the emp_emp_id_pk index?
a) You have to use extended hint syntax to be able to specify hints that apply to
tables and indexes that appear in views. The NO_MERGE hint is used to make sure
that the view is not merged into the surrounding query blocks.
SQL> @show_exec_plan_view2_hints
SQL> set echo on
SQL>

2 SELECT /*+ NO_MERGE(v2) INDEX(v2.v1.employees emp_emp_id_pk)
FULL(v2.departments) */ *
3 FROM v2
4 WHERE department_id = 30;
Explained.
SQL>
SQL> select * from table(DBMS_XPLAN.DISPLAY(null,null,'ALL'));
PLAN_TABLE_OUTPUT
--------------------------------------------------------------------
--------------------------------------------------------------------
----------------------------------------------------------------
n s e
i ce
--------------------------------------------------------------------
b l el
---------------------------
fer a
a n s
-tr
--------------------------------------------------------------------
n
o
an
---------------------------
78 | 5 (0)| 00:00:01 |
h s
a deฺ
| 1 | VIEW | V2
e )
s Gui | 3 |
r ฺ
mi dent
78 | 5 (0)| 00:00:01 |
| 2 | NESTED LOOPS
p ฺ
g Stu
| | |
| | |
d
|
57 |
3 |
5
NESTED LOOPS
b i a@ this
(0)| 00:00:01 |
| | 3 |
|* 4 |
i a (ja use
TABLE ACCESS FULL | DEPARTMENTS | 1 |
8 | 3
A b to
(0)| 00:00:01 |
o|n io
|* 5 |
1
INDEX RANGE SCAN
(0)| 00:00:01 |
| EMP_EMP_ID_PK | 50 |
nt
e A |* 6 | TABLE ACCESS BY INDEX ROWID| EMPLOYEES | 3 |
J o s 33 | 2 (0)| 00:00:01 |
--------------------------------------------------------------------
---------------------------

-------------------------------------------------------------
1 - SEL$335DD26A / V2@SEL$1
2 - SEL$335DD26A
4 - SEL$335DD26A / DEPARTMENTS@SEL$2
5 - SEL$335DD26A / EMPLOYEES@SEL$3
6 - SEL$335DD26A / EMPLOYEES@SEL$3

---------------------------------------------------
4 - filter("DEPARTMENTS"."DEPARTMENT_ID"=30)
5 - access("EMPLOYEE_ID"<150)
6 - filter("DEPARTMENT_ID"=30)

-----------------------------------------------------------
1 - "V2"."EMPLOYEE_ID"[NUMBER,22],
"V2"."DEPARTMENT_ID"[NUMBER,22]
"EMPLOYEE_ID"[NUMBER,22]
"EMPLOYEES".ROWID[ROWID,10], "EMPLOYEE_ID"[NUMBER,22]
4 - "DEPARTMENTS"."DEPARTMENT_ID"[NUMBER,22]
5 - "EMPLOYEES".ROWID[ROWID,10], "EMPLOYEE_ID"[NUMBER,22]
39 rows selected.
SQL>
--------------------------------------------------------------
n s e
i ce
set echo on
b l el
explain plan for
fer a
SELECT /*+ NO_MERGE(v2) INDEX(v2.v1.employees emp_emp_id_pk)
a n s
FULL(v2.departments) */ *
o n -tr
an
FROM v2
WHERE department_id = 30;
h s
a deฺ
)
s Gui
select * from table(DBMS_XPLAN.DISPLAY(null,null,'ALL'));
e
r ฺ
i ent
p ฺ m
g andSclean
12) Exit from your SQL*Plus session
d t udup your environment by executing the
hr_hint_cleanup.sh
b i a@script.
t h is
SQL> exit; (ja se Database 11g Enterprise Edition Release
i a to
b - Production
Disconnected from u
Oracle
o A
ni the Partitioning, Oracle Label Security, OLAP, Data Mining
11.1.0.6.0
t oWith
e An and Real Application Testing options
Jos
[oracle@edrsr33p1-orcl Hints]$ ./hr_hint_cleanup.sh
Connected to:
Production
SQL>
SQL> revoke dba from hr;
revoke dba from hr
*
ERROR at line 1:
ORA-01951: ROLE 'DBA' not granted to 'HR'

SQL>
SQL> drop view hr.v1;
View dropped.
SQL>
SQL> drop view hr.v2;
View dropped.
SQL>
SQL> exit;
n s e
i ce
b l el
--------------------------------------------------------------
fe ra
#!/bin/bash
a n s
o n -tr
an
h s
a deฺ
e )
s Gui
r ฺ
i ent
p ฺ m d
export
d g S t u
b i a@ this
sqlplus b/ia
(ja u@hr_hint_cleanup.sql
se
A to
as sysdba
ni o
nto--------------------------------------------------------------
e A
J o s set echo on
revoke dba from hr;
drop view hr.v1;
drop view hr.v2;
exit;
13) In this third case, you investigate how to influence optimizer’s joins. From your
terminal session, execute the sh_hint_setup.sh script.
[oracle@edrsr33p1-orcl Hints]$ ./sh_hint_setup.sh

Connected to:
Production
SQL>
SQL> alter user sh identified by sh acccount unlock;
alter user sh identified by sh acccount unlock
*
ERROR at line 1:
ORA-00922: missing or invalid option
SQL>
SQL> grant dba to sh;
n s e
i ce
Grant succeeded.
b l el
SQL>
fer a
SQL> connect sh/sh
a n s
Connected.
o n -tr
an
SQL>
SQL> exec dbms_stats.delete_schema_stats('SH');
h s
a deฺ
e )
s Gui
r ฺ
i ent
SQL>
p ฺ m d
SQL> exec
d g S t u
b i a@ this
dbms_stats.set_table_stats('SH','SALES',null,null,null,10,5);
PL/SQL procedure
u se completed.
SQL> A
b to
n o
i exit;
ntoSQL>
e A Disconnected from Oracle Database 11g Enterprise Edition Release
J o s 11.1.0.6.0 - Production
--------------------------------------------------------------
#!/bin/bash
export
sqlplus / as sysdba @sh_hint_setup.sql

--------------------------------------------------------------
set echo on
alter user sh identified by sh acccount unlock;

grant dba to sh;
connect sh/sh
exec dbms_stats.delete_schema_stats('SH');
exec dbms_stats.set_table_stats('SH','SALES',null,null,null,10,5);
exit;
n s e
i ce
14) From your terminal session, connect as the SH user in the SQL*Plus session. After b l el
you are connected, ensure that you flush your SGA content using the f e ra
flush_sga.sql script, and set some important session parameters a s the
nusing
n t r
- from your
set_sh_session.sql script. Ensure that you do not disconnect
o
established SQL*Plus session.
s an
h a
sh/sh ide
ฺ
)
rฺes nt GonuTue Apr 1 19:38:35 2008
[oracle@edrsr33p1-orcl Hints]$ sqlplus
SQL*Plus: Release 11.1.0.6.0 m - iProduction

d g pฺ tude
i a @ his S All rights reserved.
( j ab se t
Connectedia
b to:
t o u
i o
Oracle A Database 11g Enterprise Edition Release 11.1.0.6.0 -
n the Partitioning, Oracle Label Security, OLAP, Data Mining
toWith
Production
n
A and Real Application Testing options
e
Jos SQL> @flush_sga
SQL>
System altered.
SQL>
System altered.
SQL>
SQL> @set_sh_session
SQL>
SQL>
SQL>
SQL> set timing on

SQL>
Session altered.
Elapsed: 00:00:00.00
SQL>
--------------------------------------------------------------
set echo on
n s e
--------------------------------------------------------------
i ce
b l el
set linesize 200
fe r a
set pagesize 200
a n s
o n -tr
an
set timing on
alter session set optimizer_dynamic_sampling=0;

h s
a deฺ
e )
s Gui
r ฺ
15) From your SQL*Plus session, execute
p ฺ mithe following
d e nt query and determine its
execution plan:
select count(*) from@
dg S tu
b i a sales s,
t h is customers c where s.cust_id=c.cust_id;
a
SQL> set iecho
(jaon use
SQL> @exec_and_show_sh_exec_plan
SQL> A
b to
o
ni select count(*) from sales s, customers c where
t oSQL>
e An s.cust_id=c.cust_id;
Jos COUNT(*)
----------
918843
Elapsed: 00:00:03.59
SQL>
SQL> select * from
PLAN_TABLE_OUTPUT
--------------------------------------------------------------------
--------------------------------------------------------------------
----------------------------------------------------------------
SQL_ID 88r79fy8nphrc, child number 0
-------------------------------------
select count(*) from sales s, customers c where s.cust_id=c.cust_id

--------------------------------------------------------------------
----------------------------------
(%CPU)| Time | Pstart| Pstop |
--------------------------------------------------------------------
----------------------------------
(100)| | | |
| 1 | SORT AGGREGATE | | 1 | 26 |
| | | |
| 2 | NESTED LOOPS | | 121K| 3081K| 3
(0)| 00:00:01 | | |
| 3 | PARTITION RANGE ALL| | 10 | 130 | 3
(0)| 00:00:01 | 1 | 28 |
| 4 | TABLE ACCESS FULL | SALES | 10 | 130 | 3
(0)| 00:00:01 | 1 | 28 |
n s e
|* 5 | INDEX UNIQUE SCAN | CUSTOMERS_PK | 12138 | 154K| 0
i ce
(0)| | | |
b l el
--------------------------------------------------------------------
----------------------------------
fe r a
a n s
o n -tr
an
---------------------------------------------------
h s
a deฺ
e )
s Gui
r ฺ
22 rows selected.
p ฺ mi dent
Elapsed: 00:00:00.15@d
g Stu
SQL> b i a t h is
i a (ja use
b to
--------------------------------------------------------------
A
o
ni echo on
ntoset
e A
J o s select count(*) from sales s, customers c where s.cust_id=c.cust_id;
16) Before you continue, ensure that you flush the content of your SGA to avoid caching
issues later. Use the flush_sga.sql script.
SQL> @flush_sga
SQL> set echo on
SQL>
System altered.
Elapsed: 00:00:00.12
SQL>
System altered.

Elapsed: 00:00:00.18
SQL>
--------------------------------------------------------------
set echo on
17) Using hints only, how would you enhance the performance of the same query you
executed in step 15? Make sure you verify your implementation.
a) In step 15, the optimizer chose a NESTED LOOPS join. You can try to force a
n s e
hash join instead, using the LEADING and USE_HASH hints.
i ce
b l el
SQL> @exec_and_show_sh_hint_exec_plan
SQL> set echo on fe r a
SQL>
a n s
o n -tr
SQL> select /*+ LEADING(c s) USE_HASH(s) */ count(*) from sales s,
customers c where s.cust_id=c.cust_id;
a n
COUNT(*)
h s
a deฺ
----------
e )
s Gui
r ฺ
mi dent
918843
p ฺ
g Stu
Elapsed: 00:00:00.40
d
SQL>
b i a@ this
(ja use
SQL> select * from
b i a to
o A
ni
PLAN_TABLE_OUTPUT
nto--------------------------------------------------------------------
e A --------------------------------------------------------------------
J o s ----------------------------------------------------------------
SQL_ID 9k8wcn7ckb7c1, child number 0
-------------------------------------
select /*+ LEADING(c s) USE_HASH(s) */ count(*) from sales s,
customers
c where s.cust_id=c.cust_id
--------------------------------------------------------------------
-------------------------------------------
|TempSpc| Cost (%CPU)| Time | Pstart| Pstop |
--------------------------------------------------------------------
-------------------------------------------
| 182 (100)| | | |
| 1 | SORT AGGREGATE | | 1 | 26 |
| | | | |
|* 2 | HASH JOIN | | 121K| 3081K|
2968K| 182 (2)| 00:00:02 | | |

| 3 | INDEX FAST FULL SCAN| CUSTOMERS_PK | 121K| 1540K|
| 9 (0)| 00:00:01 | | |
| 4 | PARTITION RANGE ALL | | 10 | 130 |
| 3 (0)| 00:00:01 | 1 | 28 |
| 5 | TABLE ACCESS FULL | SALES | 10 | 130 |
| 3 (0)| 00:00:01 | 1 | 28 |
--------------------------------------------------------------------
-------------------------------------------

---------------------------------------------------
23 rows selected.
n s e
i ce
Elapsed: 00:00:00.13
b l el
SQL>
fer a
--------------------------------------------------------------
a n s
o n -tr
a nfrom sales s,
set echo on
h
select /*+ LEADING(c s) USE_HASH(s) */ count(*) s
a deฺ
customers c where s.cust_id=c.cust_id;
e )
s Gui
r ฺ
ฺ mi dent
p
d g Stu
i a@session
18) Exit from your SQL*Plus
b t h s clean up your environment by executing the
iand
a (ja usescript.
sh_hint_cleanup.sh
i
A b to
ni o
SQL> exit
n t oDisconnected from Oracle Database 11g Enterprise Edition Release
e A 11.1.0.6.0 - Production
Jos
[oracle@edrsr33p1-orcl Hints]$ ./sh_hint_cleanup.sh
Connected to:
Production
SQL> SQL> SQL> SQL> SQL> SQL> SQL> SQL> SQL> SQL> SQL> Connected.
SQL> SQL>
User altered.
SQL> SQL> revoke dba from sh

*
ERROR at line 1:
ORA-01951: ROLE 'DBA' not granted to 'SH'
SQL> SQL> SQL> Rem

SQL> Rem $Header: sh_main.sql 06-mar-2008.15:00:45 cbauwens Exp $

SQL> Rem
SQL> Rem sh_main.sql
SQL> Rem
...
SQL> Rem
SQL>
SQL> SET ECHO OFF
specify password for SH as parameter 1:

n s e
i ce
specify default tablespace for SH as parameter 2:
b l el
specify temporary tablespace for SH as parameter 3:
fer a
a n s
specify password for SYS as parameter 4:
o n -tr
n
specify directory path for the data files asaparameter 5:
a s e ฺ
h as idparameter
writeable directory path for the log )files
ฺ e s u 6:
7:i
specify version as parameterฺm
r
e n tG
d gp Stud
Session altered.bia
@ his
j a e t
(
ia to u s
b
i oA
User dropped.
n
ntoold 1: CREATE USER sh IDENTIFIED BY &pass
e A
J o s new 1: CREATE USER sh IDENTIFIED BY sh
User created.
old 1: ALTER USER sh DEFAULT TABLESPACE &tbs

new 1: ALTER USER sh DEFAULT TABLESPACE example
old 2: QUOTA UNLIMITED ON &tbs
new 2: QUOTA UNLIMITED ON example
User altered.
old 1: ALTER USER sh TEMPORARY TABLESPACE &ttbs

new 1: ALTER USER sh TEMPORARY TABLESPACE temp
User altered.
Grant succeeded.
...

Grant succeeded.

Connected.
Grant succeeded.
old 1: CREATE OR REPLACE DIRECTORY data_file_dir AS '&data_dir'

new 1: CREATE OR REPLACE DIRECTORY data_file_dir AS
'/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/'
Directory created.
n s e
i ce
old 1: CREATE OR REPLACE DIRECTORY log_file_dir AS '&log_dir'
b l el
new 1: CREATE OR REPLACE DIRECTORY log_file_dir AS '/home/oracle/'
fer a
Directory created.
a n s
n -tr
o
Grant succeeded.
s an
) h a deฺ
r ฺ e s Gui
mi dent
Grant succeeded.
p ฺ
g Stu
d
Grant succeeded.
b i a@ this
Connected. (ja se
i a
baltered.to u
i o A
Session
nton
e A
J o s Session altered.
Table created.
...
Comment created.
Creating OLAP metadata ...

<<<<< CREATE CWMLite Metadata for the Sales History Schema >>>>>
-
…
-

Commit complete.
gathering statistics ...


n s e
i ce
b l
19) With this fourth case, you study the influence of the INDEX and AND_EQUAL hints.el
fer a
From your terminal session, execute the sh_hint_index_setup.sh script to set
up the environment for this lab. a n s
o n -tr
an
[oracle@edrsr33p1-orcl Hints]$ ./sh_hint_index_setup.sh

h s
a on Tue ฺ 1 20:16:22 2008
eApr
) i d
Copyright (c) 1982, 2007, Oracle.i r ฺesAllnrights
t Gu reserved.
g p ฺm ude
d S t
@ s
thi Edition Release 11.1.0.6.0 -
ia Enterprise
Connected to:
Oracle Database b
Production ( j a 11g
e
b a to us Oracle Label Security, OLAP, Data Mining
iPartitioning,
A Application Testing options
With the
andoReal
i
to n
A n SQL>
o s e SQL> alter user sh identified by sh acccount unlock;

J alter user sh identified by sh acccount unlock
*
ERROR at line 1:
ORA-00922: missing or invalid option
SQL>
SQL> grant dba to sh;
Grant succeeded.
SQL>
SQL> exit;
--------------------------------------------------------------

#!/bin/bash
export
sqlplus / as sysdba @sh_hint_index_setup.sql
--------------------------------------------------------------
n s e
ice
set echo on
b l el
alter user sh identified by sh acccount unlock;
fe ra
a n s
grant dba to sh;
o n -tr
exit;
s an
) h a deฺ
20) From your terminal session, connect to rtheฺ e s Gsession
SQL*Plus ui as the SH user. After
ฺ mi scripts
you are connected, execute the following
p d e ntot further set up your environment
for this lab. Ensure that you d g connected
stay S tu to your session throughout this case.
b i a@ Hints]$
t h is sqlplus sh/sh
(ja use
SQL*Plus: b i a to11.1.0.6.0 - Production on Tue Apr 1 20:16:28 2008

Release
io A
t o n
A n
e
Jos Connected to:
Production
SQL> @drop_index_customers
SQL>
SQL> @dait
SQL>
SQL> drop index CUSTOMERS_YOB_BIX;
Index dropped.
SQL> drop index CUSTOMERS_MARITAL_BIX;
Index dropped.
SQL> drop index CUSTOMERS_GENDER_BIX;

Index dropped.
SQL>
SQL>
SQL> @create_cust_indexes
SQL> set echo on
SQL>
SQL> CREATE INDEX CUST_CUST_GENDER_idx
2 ON CUSTOMERS(CUST_GENDER)
Index created.
SQL>
SQL> CREATE INDEX CUST_CUST_POSTAL_CODE_idx
2 ON CUSTOMERS(CUST_POSTAL_CODE)
n s e
i ce
b l el
Index created.
fer a
SQL>
a n s
SQL> CREATE INDEX CUST_CUST_CREDIT_LIMIT_idx
o n -tr
an
2 ON CUSTOMERS(CUST_CREDIT_LIMIT)
h s
a deฺ
Index created.
e )
s Gui
r ฺ
SQL>
p ฺ mi dent
d g Stu
b i a@ this
--------------------------------------------------------------
REM drop (
a ja indexes
all
u seindexes
on CUSTOMERS table
REM b i
does not
to
touch associated with constraints
REM A==================================================
o n io
nt set termout off
e A
J o s store
save
set sqlplus_settings replace
buffer.sql replace
set heading off verify off autotrace off feedback off
spool dait.sql
SELECT 'drop index '||i.index_name||';'

FROM user_indexes i
WHERE i.table_name = 'CUSTOMERS'
AND NOT EXISTS
(SELECT 'x'
FROM user_constraints c
WHERE c.index_name = i.index_name
AND c.table_name = i.table_name
AND c.status = 'ENABLED');
spool off
get buffer.sql nolist

@sqlplus_settings
set termout on

set echo on
@dait
--------------------------------------------------------------
set echo on
CREATE INDEX CUST_CUST_GENDER_idx

ON CUSTOMERS(CUST_GENDER)
NOLOGGING COMPUTE STATISTICS;
CREATE INDEX CUST_CUST_POSTAL_CODE_idx

ON CUSTOMERS(CUST_POSTAL_CODE)
n s e
i ce
CREATE INDEX CUST_CUST_CREDIT_LIMIT_idx
b l el
ON CUSTOMERS(CUST_CREDIT_LIMIT)
fer a
a n s
n - tr
o
21) Determine the execution plan for the following query:
SELECT
s an
c.*
) h a deฺ
FROM customers c
r ฺ e s Gui
WHERE cust_gender
p ฺ mi dent
= 'M'
AND g Stu
d
AND
a@ this
cust_credit_limit = 10000;
b i
a (ja 200
SQL> set autotrace
SQL> set ilinesize
sepagesize 1000
traceonly
u
A b to
o
SQL> @index
ni set echo on
n t oSQL>
e A SQL>
Jos
SQL> SELECT
2 c.*
3 FROM customers c
7 /
6 rows selected.
Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
-----------
Time |
--------------------------------------------------------------------
-----------

00:00:05 |
00:00:05 |
--------------------------------------------------------------------
-----------

---------------------------------------------------
"CUST_CREDIT_LIMIT"=10000 AND "CUST_GENDER"='M')
Statistics
----------------------------------------------------------
n s e
1 recursive calls
i ce
0 db block gets
b l el
0 physical reads
fe ra
0 redo size
a n s
o n -tr
an
0 sorts (memory)
h s
a deฺ
0 sorts (disk)
e )
s Gui
i ent
p ฺ m
SQL>
d g Stud
b i a@ this
--------------------------------------------------------------
i a (ja use
set echo
A b on
to
n i o
ntoSELECT
e A c.*
J o s FROM customers c
AND cust_credit_limit = 10000
/
22) Try to get a better execution plan using the INDEX hint for the same query you
investigated in step 21. Which index is best suited?
a) The CUST_CUST_GENDER_IDX index is the best one for this query. Note that
using the INDEX hint without specifying any index leads the optimizer to use the
CUST_CUST_GENDER_IDX index.
SQL> @index_hint
SQL> set echo on
SQL>
SQL> SELECT /*+ INDEX (c &indexname) */
2 c.*
3 FROM customers c

7 /
Enter value for indexname: CUST_CUST_CREDIT_LIMIT_IDX
old 1: SELECT /*+ INDEX (c &indexname) */
new 1: SELECT /*+ INDEX (c CUST_CUST_CREDIT_LIMIT_IDX) */
6 rows selected.
Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
--------------------------------------
n s e
i ce
b l el
--------------------------------------------------------------------
--------------------------------------
fer a
a n s |
6 | 1080 | 1074 (1)| 00:00:11 |
o n -tr
an
6 | 1080 | 1074 (1)| 00:00:11 |
h s
a deฺ
| CUST_CUST_CREDIT_LIMIT_IDX |
6938 | | 14 (0)| 00:00:01 |
e )
s Gui
r ฺ
mi dent
--------------------------------------------------------------------
p ฺ
--------------------------------------
g Stuby operation id):
d
b i a@ this
(identified
---------------------------------------------------
i a (ja use
b to
1 - filter(TO_NUMBER("CUST_POSTAL_CODE")=40804
A
AND
o n io2 - access("CUST_CREDIT_LIMIT"=10000)
"CUST_GENDER"='M')
nt
e A
J o s Statistics
----------------------------------------------------------
1 recursive calls
0 db block gets
13 physical reads
0 redo size
0 sorts (memory)
0 sorts (disk)
6 rows processed
SQL>
SQL> @index_hint
SQL> set echo on
SQL>
2 c.*

3 FROM customers c
7 /
Enter value for indexname: CUST_CUST_GENDER_IDX

new 1: SELECT /*+ INDEX (c CUST_CUST_GENDER_IDX) */
6 rows selected.
Execution Plan
----------------------------------------------------------
n s e
--------------------------------------------------------------------
i ce
--------------------------------
b l el
| Id | Operation
| Name | Rows |
fer a
n s
--------------------------------------------------------------------
a
--------------------------------
o n -tr
an
1080 | 1164 (1)| 00:00:12 |
h s
a deฺ
|* 1 | TABLE ACCESS BY INDEX ROWID| CUSTOMERS | 6 |
1080 | 1164 (1)| 00:00:12 |
e )
s Gui
r ฺ
mi dent
|* 2 | INDEX RANGE SCAN | CUST_CUST_GENDER_IDX | 27750 |
| 51 (0)| 00:00:01 |
p ฺ
g Stu
d
--------------------------------------------------------------------
b i a@ this
--------------------------------

b to
---------------------------------------------------
A
o
ni 1 - filter(TO_NUMBER("CUST_POSTAL_CODE")=40804 AND
nto"CUST_CREDIT_LIMIT"=10000)
e A
J o s 2 - access("CUST_GENDER"='M')
Statistics
----------------------------------------------------------
1 recursive calls
0 db block gets
68 physical reads
0 redo size
0 sorts (memory)
0 sorts (disk)
6 rows processed
SQL>
SQL> @index_hint
SQL> set echo on
SQL>

2 c.*
3 FROM customers c
7 /
Enter value for indexname: CUST_CUST_POSTAL_CODE_IDX
new 1: SELECT /*+ INDEX (c CUST_CUST_POSTAL_CODE_IDX) */
6 rows selected.
Execution Plan
----------------------------------------------------------
n s e
i ce
b l el
--------------------------------------------------------------------
-------------------------------------
fer a
a n s |
n - tr
o
an
--------------------------------------------------------------------
-------------------------------------
h s
a deฺ |
6 | 1080 | 218 (1)| 00:00:03 |
e )
s Gui
r ฺ
mi dent
6 | 1080 | 218 ฺ
(1)| 00:00:03 |
p
g Stu
|* 2 |
d
INDEX FULL SCAN | CUST_CUST_POSTAL_CODE_IDX |
89 | | 134
b i a@ this
(1)| 00:00:02 |
--------------------------------------------------------------------
i a (ja use
-------------------------------------
Ab to
o n io
---------------------------------------------------
nt
e A
J o s 1 - filter("CUST_CREDIT_LIMIT"=10000 AND "CUST_GENDER"='M')
Statistics
----------------------------------------------------------
1 recursive calls
0 db block gets
251 consistent gets
132 physical reads
0 redo size
0 sorts (memory)
0 sorts (disk)
6 rows processed
SQL>
SQL> @index_hint
SQL> set echo on

SQL>
2 c.*
3 FROM customers c

7 /
Enter value for indexname:
new 1: SELECT /*+ INDEX (c ) */
6 rows selected.
Execution Plan
n s e
----------------------------------------------------------
i ce
b l el
fer
-------------------------------------------------------------------- a
-------------------------------------
a n s
n - tr |
o
an
--------------------------------------------------------------------
-------------------------------------
h s
a deฺ
e )
s Gui |
r ฺ
mi dent
6 | 1080 | 218 (1)| 00:00:03 |
p ฺ
|* 1 | TABLE ACCESS BY INDEX ROWID| CUSTOMERS
g Stu
|
6 | 1080 | 218
d
(1)| 00:00:03 |
|* 2 |
89 | | 134b a@ this
INDEX FULL SCAN
i(1)| 00:00:02 |
| CUST_CUST_POSTAL_CODE_IDX |
i a (ja use
--------------------------------------------------------------------
b to
-------------------------------------
A
o n io
nt
e A ---------------------------------------------------
J o s 1 - filter("CUST_CREDIT_LIMIT"=10000 AND "CUST_GENDER"='M')

Statistics
----------------------------------------------------------
1 recursive calls
0 db block gets
251 consistent gets
0 physical reads
0 redo size
0 sorts (memory)
0 sorts (disk)
6 rows processed
SQL>

--------------------------------------------------------------
set echo on
SELECT /*+ INDEX (c &indexname) */

c.*
FROM customers c
/
23) Investigate the same query, but this time using the AND_EQUAL hint. Can you find a
better execution plan with this hint?
n s e
a) Using the AND_EQUAL hint allows the optimizer to explicitly choose an
i ce
execution plan that uses an access path that merges the scans on several single-
b l el
column indexes. However, no combination is better than the
fe r a
CUST_CUST_GENDER_IDX index.
a n s
n - tr
SQL> @and_equal_hint
o
SQL> set echo on
s an
SQL>
) h a deฺ
s Gui
SQL> SELECT /*+ AND_EQUAL (c &index_name1, &index_name2) */
2 c.*
r ฺ e
3 FROM customers c
p ฺ mi dent
4 WHERE cust_gender
5 AND d g Stu
= 'M'
6 AND i a@ this
cust_credit_limit = 10000
b
7 /
i a (ja use
Enter value for index_name1: CUST_CUST_CREDIT_LIMIT_IDX
Ab to
Enter value for index_name2: CUST_CUST_GENDER_IDX
o n
oldio 1: SELECT /*+ AND_EQUAL (c &index_name1, &index_name2) */
nt new 1: SELECT /*+ AND_EQUAL (c CUST_CUST_CREDIT_LIMIT_IDX,
e A CUST_CUST_GENDER_IDX) */
J o s
6 rows selected.
Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
--------------------------------------
--------------------------------------------------------------------
--------------------------------------
6 | 1080 | 1416 (1)| 00:00:15 |
6 | 1080 | 1416 (1)| 00:00:15 |
| 2 | AND-EQUAL | |
| | | |

|* 3 | INDEX RANGE SCAN | CUST_CUST_CREDIT_LIMIT_IDX |
6938 | | 14 (0)| 00:00:01 |
|* 4 | INDEX RANGE SCAN | CUST_CUST_GENDER_IDX |
27750 | | 51 (0)| 00:00:01 |
--------------------------------------------------------------------
--------------------------------------

---------------------------------------------------
"CUST_CREDIT_LIMIT"=10000 AND
"CUST_GENDER"='M')
n s e
i ce
Statistics
b l el
----------------------------------------------------------
1 recursive calls
fer a
0 db block gets
a n s
o n -tr
an
0 physical reads
0 redo size
h
2570 bytes sent via SQL*Net to clients
a deฺ
)
s Gui
e
r ฺ
mi dent
0 sorts (memory)
p ฺ
g Stu
d
0 sorts (disk)
b i a@ this
6 rows processed
SQL>
i a (ja use
b echo onto
SQL> @and_equal_hint
A
i o
SQL> set
n
t oSQL>
n SQL> SELECT /*+ AND_EQUAL (c &index_name1, &index_name2) */
e A
Jos
2 c.*
3 FROM customers c
7 /
Enter value for index_name1: CUST_CUST_CREDIT_LIMIT_IDX
Enter value for index_name2: CUST_CUST_POSTAL_CODE_IDX
old 1: SELECT /*+ AND_EQUAL (c &index_name1, &index_name2) */
new 1: SELECT /*+ AND_EQUAL (c CUST_CUST_CREDIT_LIMIT_IDX,
CUST_CUST_POSTAL_CODE_IDX) */
6 rows selected.
Execution Plan
----------------------------------------------------------
--------------------------------------------------------------------
-------------------------------------

--------------------------------------------------------------------
-------------------------------------
6 | 1080 | 218 (1)| 00:00:03 |

6 | 1080 | 218 (1)| 00:00:03 |
|* 2 | INDEX FULL SCAN | CUST_CUST_POSTAL_CODE_IDX |
89 | | 134 (1)| 00:00:02 |
--------------------------------------------------------------------
-------------------------------------

---------------------------------------------------
n s e
1 - filter("CUST_CREDIT_LIMIT"=10000 AND "CUST_GENDER"='M')
i ce
b l el
fer a
Statistics
a n s
-
----------------------------------------------------------
n tr
o
an
1 recursive calls
0 db block gets
251 consistent gets
h s
a deฺ
0 physical reads
e )
s Gui
r ฺ
mi dent
0 redo size
p ฺ
g Stu
d
b i a@ this
0 sorts (memory)
i a (ja use
0 sorts (disk)
Ab to
6 rows processed
o n io
SQL>
nt
e A
J o s --------------------------------------------------------------
set echo on
SELECT /*+ AND_EQUAL (c &index_name1, &index_name2) */

c.*
FROM customers c
/
24) Exit from your SQL*Plus session and clean up your environment by executing the
sh_hint_cleanup.sh script.
SQL> exit

[oracle@edrsr33p1-orcl Hints]$ ./sh_hint_cleanup.sh

Connected to:
Production
SQL> SQL> SQL> SQL> SQL> SQL> SQL> SQL> SQL> SQL> SQL> Connected.
SQL> SQL>
n s e
User altered.
i ce
b l el
SQL> SQL> revoke dba from sh
*
fer a
ERROR at line 1:
a n s
ORA-01951: ROLE 'DBA' not granted to 'SH'
o n -tr
s an
SQL> SQL> SQL> Rem
) h a deฺ cbauwens Exp $
r ฺ e s Gui
SQL> Rem $Header: sh_main.sql 06-mar-2008.15:00:45
SQL> Rem
SQL> Rem sh_main.sql ฺ m i ent
SQL> Rem
d g p t u d
...
a@ this S
SQL> Rem b i
SQL>
i a (jaOFF use
SQL> SET
A b ECHO
to
ni o
ntospecify password for SH as parameter 1:
e A
J o s specify default tablespace for SH as parameter 2:
specify directory path for the data files as parameter 5:
writeable directory path for the log files as parameter 6:
specify version as parameter 7:
Session altered.
User dropped.
old 1: CREATE USER sh IDENTIFIED BY &pass

new 1: CREATE USER sh IDENTIFIED BY sh

User created.

User altered.

User altered.
Grant succeeded.
n s e
i ce
b l el
...
fer a
a n s
Grant succeeded.
o n -tr
PL/SQL procedure successfully completed.as
an
) h i d eฺ
Connected.
i r ฺes nt Gu
g p ฺm ude
Grant succeeded.
d S t
old 1: CREATE b OR
@ s
iaREPLACEthiDIRECTORY data_file_dir AS '&data_dir'
new (
1: CREATEj a OR se DIRECTORY data_file_dir AS
REPLACE
b i a to u
o A
ni
toDirectory created.
A n
o s e old 1: CREATE OR REPLACE DIRECTORY log_file_dir AS '&log_dir'
J new 1: CREATE OR REPLACE DIRECTORY log_file_dir AS '/home/oracle/'
Directory created.
Grant succeeded.
Grant succeeded.
Grant succeeded.
Connected.
Session altered.
Session altered.

Table created.
...
Comment created.

…
n s e
i ce
b l el
Commit complete.
fer a
a n s
o n -tr
an
PL/SQL procedure successfully completed.as

) h i d eฺ
i r ฺ es t Gu
g p ฺmcompleted.
u d en
dfrom Oracle
S t Database 11g Enterprise Edition
@ s
thi Label Security, OLAP, Data Mining
SQL> SQL> Disconnected
Release 11.1.0.6.0
a b ia- Production
j e
b
and Real a ( to usTesting options
iApplication
Oracle
i oA
n
Hints]$
nto--------------------------------------------------------------
e A
J o s #!/bin/bash
cd /home/oracle/solutions/SQL_Access_Advisor/sh
export
cp * $ORACLE_HOME/demo/schema/sales_history
SET ECHO ON
SET FEEDBACK 1
SET NUMWIDTH 10
SET LINESIZE 8000

SET TRIMSPOOL ON
SET TAB OFF
SET PAGESIZE 100
SET LONG 1000
CONNECT / AS SYSDBA
revoke dba from sh;
@sh_main sh example temp oracle

/home/oracle/ v3
exit;
n s e
i ce
FIN!
b l el
ra
25) In the fifth case, you retrieve the first rows of your query as fast as possible. Connect
fe
n s
to the SQL*Plus session as the SYS user and ensure that you set the following
a
SQL*Plus environment variables:
o n -tr
timing on
s an
autotrace on
) h a deฺ
pagesize 200
r ฺ e s Gui
linesize 200
ฺ m i ent
d p
[oracle@edrsr33p1-orclgHints]$ d
tusqlplus / as sysdba
a@ S
is - Production on Tue Apr 29
b i t h
(ja use
i
21:02:19 2008
b a to
i o A
t o n
A n
o s e
J Connected to:
Production
With the Partitioning, Oracle Label Security, OLAP, Data
Mining
SQL> set timing on autotrace on pagesize 200 linesize 200

SQL>
26) From the same SQL*Plus session, execute the following query. Based on the fact that
you want to retrieve the first 10 rows as fast as possible, what do you observe?
SELECT employee_id, department_name
FROM hr.employees e, hr.departments d
WHERE e.department_id = d.department_id;
a) It takes a bit of time before the result starts to be printed on the screen. This is
because a merge join is used. It needs to sort all data before producing rows.

SQL> SELECT employee_id, department_name
FROM hr.employees e, hr.departments d
WHERE e.department_id = d.department_id;
2 3
EMPLOYEE_ID DEPARTMENT_NAME
----------- ------------------------------
200 Administration
201 Marketing
202 Marketing
114 Purchasing
115 Purchasing
116 Purchasing
117 Purchasing
…
113 Finance n s e
i ce
el
205 Accounting
206 Accounting
a b l
fer
106 rows selected.
a n s
o n -tr
an
Elapsed: 00:00:00.09
Execution Plan h s
a deฺ
e )
s Gui
Plan hash value: 1343509718 i r ฺ
----------------------------------------------------------
p ฺ m dent
d g Stu
a@ this
--------------------------------------------------------------
b i
------------------------------
a
a (j (%CPU)|
| Id | Operation
Bytes |biCost u seTime | Name | Rows |
A to |
ni o
--------------------------------------------------------------
nto------------------------------
e A | 0 | SELECT STATEMENT | | 106 |
J o s 2438 | 6 (17)| 00:00:01 |

| 1 | MERGE JOIN | | 106 |
2438 | 6 (17)| 00:00:01 |
| 2 | TABLE ACCESS BY INDEX ROWID| DEPARTMENTS | 27 |
432 | 2 (0)| 00:00:01 |
| 3 | INDEX FULL SCAN | DEPT_ID_PK | 27 |
| 1 (0)| 00:00:01 |
|* 4 | SORT JOIN | | 107 |
749 | 4 (25)| 00:00:01 |
| 5 | TABLE ACCESS FULL | EMPLOYEES | 107 |
749 | 3 (0)| 00:00:01 |
--------------------------------------------------------------
------------------------------

---------------------------------------------------
4 - access("E"."DEPARTMENT_ID"="D"."DEPARTMENT_ID")
filter("E"."DEPARTMENT_ID"="D"."DEPARTMENT_ID")

Statistics
----------------------------------------------------------
0 recursive calls
0 db block gets
19 consistent gets
0 physical reads
0 redo size
1 sorts (memory)
0 sorts (disk)
106 rows processed n s e
i ce
SQL>
b l el
fe r a
a n s
27) Using a hint, how can you ensure that the previous query starts fetching rows faster?
Test your solution.
o n -tr
a) Using the FIRST_ROWS(10) hint, a nested loop is chosen
a n by the optimizer. It is
faster to retrieve the first rows.
h a s eฺ
SQL> SELECT /*+ FIRST_ROWS(10) e*/ s u id department_name
)employee_id,
ฺ
ir entd G
FROM hr.employees e, hr.departments
=p ฺ m
WHERE e.department_id g
d tud
d.department_id;
S
2 3
b i a@ this
EMPLOYEE_ID DEPARTMENT_NAME
a se
-----------(j------------------------------
i a to u
b200 Administration
i o A 201 Marketing
nton 202 Marketing
e A 114 Purchasing
J o s 115 Purchasing
116 Purchasing
…
112 Finance
113 Finance
205 Accounting
206 Accounting
106 rows selected.
Elapsed: 00:00:00.02
Execution Plan
----------------------------------------------------------
--------------------------------------------------------------
------------------------------------

--------------------------------------------------------------
------------------------------------
10 | 230 | 3 (0)| 00:00:01 |
| | | |
10 | 230 | 3 (0)| 00:00:01 |
| 3 | TABLE ACCESS FULL | DEPARTMENTS |
25 | 400 | 2 (0)| 00:00:01 |
|* 4 | INDEX RANGE SCAN | EMP_DEPARTMENT_IX |
8 | | 0 (0)| 00:00:01 |
| 5 | TABLE ACCESS BY INDEX ROWID| EMPLOYEES | n s e
i ce
el
3 | 21 | 1 (0)| 00:00:01 |
--------------------------------------------------------------
a b l
------------------------------------
fer
a n s
o n -tr
an
---------------------------------------------------
h s
a deฺ
4 - access("E"."DEPARTMENT_ID"="D"."DEPARTMENT_ID")
e )
s Gui
r ฺ
p ฺ mi dent
Statistics
d g Stu
----------------------------------------------------------
b i a@ th
0 recursive is
calls
0(ja
db block se gets
i a to
b 0 physical
40 u
consistent gets
io A reads
t o n 0 redo size
A n 2435 bytes sent via SQL*Net to client
o s e 497 bytes received via SQL*Net from client
J 9 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
106 rows processed
SQL>

n s e
i ce
b l el
fer a
a n s
n - tr
o
s an
) h a deฺ
r ฺ e s Gui
p ฺ mi dent
d g Stu
b i a@ this
i a (ja use
Ab to
o n io
nt
e A
J o s

Practice 10-1: Tracing Applications
In this practice, you define a service and use it to generate traces. You then interpret
generated trace files. You can find all needed script files for this lab in your
$HOME/solutions/Application_Tracing directory.
1) Initialize your environment be executing the at_setup.sh script from a terminal
session connected as the oracle user.

[oracle@edrsr33p1-orcl Application_Tracing]$ ./at_setup.sh
SQL*Plus: Release 11.1.0.6.0 - Production on Fri Apr 4 20:25:55 2008
n s e
Connected to:
i ce
b l el
Production
fe r a
a n s
o n -tr
an
SQL>
SQL> drop user trace cascade;
drop user trace cascade
h s
a deฺ
*
e )
s Gui
r ฺ
mi dent
ERROR at line 1:
p ฺ
ORA-01918: user 'TRACE' does not exist
g Stu
d
b i a@ this
SQL>
SQL> create (
a ja trace
user
u s eidentified by trace default tablespace users
i to temp;
b tablespace
temporary
o A
ni created.
toUser
A n
o s e SQL>
J SQL> grant connect, resource, dba to trace;
Grant succeeded.
SQL>
SQL>
SQL> drop tablespace tracetbs including contents and datafiles;
drop tablespace tracetbs including contents and datafiles
*
ERROR at line 1:
ORA-00959: tablespace 'TRACETBS' does not exist
SQL>
SQL> drop tablespace tracetbs3 including contents and datafiles;
drop tablespace tracetbs3 including contents and datafiles
*
ERROR at line 1:
ORA-00959: tablespace 'TRACETBS3' does not exist

Practice 10-1: Tracing Applications (continued)
SQL>
SQL> create tablespace tracetbs
2 datafile '/u01/app/oracle/oradata/orcl/tracetbs.dbf' size 100m
3 extent management local uniform size 40k;
Tablespace created.
SQL>
SQL> create tablespace tracetbs3
2 datafile '/u01/app/oracle/oradata/orcl/tracetbs3.dbf' size 100m
3 extent management local uniform size 10m;
Tablespace created.
SQL>
n s e
SQL>
i ce
SQL> connect trace/trace
b l el
Connected.
SQL>
fer a
SQL> drop table sales purge;
a n s
drop table sales purge
n -tr
o
an
*
ERROR at line 1:
h s
a deฺ
e )
s Gui
r ฺ
i ent
SQL>
p ฺ m
d g select
SQL> create table sales as
S tu*dfrom sh.sales;
Table created. bia
@ his
j a e t
(
ia to u s
SQL>
SQL> A
b
o n io drop table sales2 purge;
SQL>
Ant drop table sales2 purge
e
Jos
*
ERROR at line 1:
SQL>
SQL> create table sales2 tablespace tracetbs as select * from
sh.sales where 1=2;
Table created.
SQL>
SQL> drop table sales3 purge;
drop table sales3 purge
*
ERROR at line 1:
SQL>

SQL> create table sales3 tablespace tracetbs3 as select * from
sh.sales where 1=2;
Table created.
SQL>
SQL> exit;
[oracle@edrsr33p1-orcl Application_Tracing]$
--------------------------------------------------------------
#!/bin/bash
n s e
i ce
cd /home/oracle/solutions/Application_Tracing
b l el
fer a
a n s
o n -tr
export
s an
h a deฺ
)
r ฺ e s Gui
ฺ mi dent
sqlplus / as sysdba @at_setup.sql
p
d g Stu
b i a@ this
--------------------------------------------------------------
set echo on (ja se

i a o u
b trace tcascade;
o A
ni user trace identified by trace default tablespace users
drop user
ntocreate
e A
J o s temporary tablespace temp;
grant connect, resource, dba to trace;
drop tablespace tracetbs including contents and datafiles;
drop tablespace tracetbs3 including contents and datafiles;
create tablespace tracetbs

datafile '/u01/app/oracle/oradata/orcl/tracetbs.dbf' size 100m
extent management local uniform size 40k;
create tablespace tracetbs3

datafile '/u01/app/oracle/oradata/orcl/tracetbs3.dbf' size 100m
extent management local uniform size 10m;
connect trace/trace
drop table sales purge;

create table sales as select * from sh.sales;
drop table sales2 purge;

create table sales2 tablespace tracetbs as select * from sh.sales

where 1=2;
drop table sales3 purge;
create table sales3 tablespace tracetbs3 as select * from sh.sales

where 1=2;
exit;
n s e
ice
is available from the tnsnames.ora file you use to connect to your database. b l el
2) Before you can use a service in your applications, you have to make sure this service
fe r a
a n s
Modify this file to make sure it references a service called TRACESERV. You can use
the add_traceserv_tns.sh script to help you in this task.
TRACESERV = o n -tr
(DESCRIPTION =
s an
) h a deฺ
(ADDRESS = (PROTOCOL = TCP)(HOST = node)(PORT = 1521))
(CONNECT_DATA =
r ฺ e s Gui
mi dent
(SERVER = DEDICATED)
(SERVICE_NAME = TRACESERV)
p ฺ
g Stu
)
d
)
b i a@ this
( j a
s eApplication_Tracing]$ ./add_traceserv_tns.sh
bia to
[oracle@edrsr33p1-orcl u Application_Tracing]$
io A
t o n
--------------------------------------------------------------
A n
o s e #!/bin/ksh
J y=`hostname`
DBNAME=orcl
sed 's/NODE/'$y'/'
/home/oracle/solutions/Application_Tracing/wrong_tnstraceserv.ora >
/home/oracle/solutions/Application_Tracing/tnstraceserv.ora
cp /u01/app/oracle/product/11.1.0/db_1/network/admin/tnsnames.ora
/u01/app/oracle/product/11.1.0/db_1/network/admin/tnsnames.ora.bak1
cat /home/oracle/solutions/Application_Tracing/tnstraceserv.ora >>
/u01/app/oracle/product/11.1.0/db_1/network/admin/tnsnames.ora
--------------------------------------------------------------
TRACESERV =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = NODE)(PORT = 1521))
(CONNECT_DATA =

(SERVER = DEDICATED)
(SERVICE_NAME = TRACESERV)
)
)
3) You now need to declare the TRACESERV service in your database. So connect to
your database instance as the SYS user using a SQL*Plus session.

[oracle@edrsr33p1-orcl Application_Tracing]$ sqlplus / as sysdba
n s e
Connected to:
i ce
b l el
Production
With the Partitioning, Oracle Label Security, OLAP, Data Mining fe r a
a n s
o n -tr
SQL>
a n
s eฺ with the same
a TRACESERV
4) In your SQL*Plus session, create a new service called
h
) i d
network name.
i r ฺes nt Gu
SQL> @add_traceserv_db
g p ฺm ude
SQL>
d S t
@ s
thi
SQL> select name from dba_services;
a b ia
j e
ia ( to us
NAME
b
----------------------------------------------------------------
i oA
SYS$BACKGROUND
n
toSYS$USERS
e An seeddataXDB
Jos
seeddata
orclXDB
orcl
6 rows selected.
SQL>
SQL> exec DBMS_SERVICE.CREATE_SERVICE('TRACESERV','TRACESERV');
SQL>
SQL> select name from dba_services;
NAME
----------------------------------------------------------------
SYS$BACKGROUND
SYS$USERS
seeddataXDB
seeddata
orclXDB

orcl
TRACESERV
7 rows selected.
SQL>
SQL>
--------------------------------------------------------------
set echo on
select name from dba_services;
exec DBMS_SERVICE.CREATE_SERVICE('TRACESERV','TRACESERV');
n s e
select name from dba_services;
i ce
b l el
5) From the same SQL*Plus session, start the TRACESERV service. fer a
a n s
SQL> @start_traceserv o n -tr
SQL> set echo on
s an
SQL>
) h a deฺ
SQL> show parameter service_names
r ฺ e s Gui
NAME
p ฺ mi deTYPE nt VALUE
d g
------------------------------------
S t u ----------- -------------------
a@ this
-----------
service_names
a b i string
SQL> ( j u s e
SQL> execia o
A b t
DBMS_SERVICE.START_SERVICE('TRACESERV');
i o
nton
e A
J o s SQL>
SQL> show parameter service_names
NAME TYPE VALUE

------------------------------------ ----------- -------------------
-----------
service_names string TRACESERV
SQL>
SQL>
--------------------------------------------------------------
set echo on
show parameter service_names
exec DBMS_SERVICE.START_SERVICE('TRACESERV');
show parameter service_names

SQL> exit;

7) At this point, open a browser window and connect to Enterprise Manager Database
Control as the SYS user. Ensure that you navigate to the Top Services page.
a) Log in to Enterprise Manager as the SYS user.
b) On the Database Home page, click the Performance tab.
c) On the Performance page, click the Top Consumers link in the Additional n s e
Monitoring links section of the page. li ce
d) On the Top Consumers page, click the Top Services tab. This takes you tora b
the
l e
f e
Top Services page.
t r a ns
8) You now execute seven workload scripts that are traced. All workload
o n - scripts run
under the TRACESERV service. Your goal is to analyze thengenerated trace files to
interpret what happens in the seven cases. From your s a
terminalฺsession, execute the
) a degeneration of statistics
htrigger ithe
run_tracep0.sh script. This script is used s to
ฺeviewed u
G within Enterprise Manager.
for your TRACESERV service so it canirbe n tfrom
ฺ m e
p of theturun_tracep0.sh
As soon as you start the execution d script, move to the
d g S
next step of this lab.
b i a@ this
i a (ja useApplication_Tracing]$ ./run_tracep0.sh
A
SQL*Plus: b Releaseto11.1.0.6.0 - Production on Fri Apr 4 20:28:40 2008
ni o
t o
e An Copyright (c) 1982, 2007, Oracle. All rights reserved.
Jos Connected to:

Production
SQL>
SQL> connect trace/trace@TRACESERV
Connected.
SQL>
SQL> alter session set tracefile_identifier='mytraceP0';
Session altered.
SQL>
SQL>
SQL> exit;

--------------------------------------------------------------
#!/bin/bash
n s e
export
i ce
b l el
fer a
sqlplus / as sysdba @run_tracep0.sql
a n s
n -tr
o
an
--------------------------------------------------------------
set echo on
h s
a deฺ
e )
s Gui
connect trace/trace@TRACESERV
r ฺ
p ฺ mi dent
alter session set tracefile_identifier='mytraceP0';
d g Stu
set termout off
b i a@ this
(ja from
select count(*)
i a u e
sdba_objects;
b
select count(*)
A o
count(*) tfrom dba_objects;
from dba_objects;
o
i count(*) from dba_objects;
select
n
ntoselect
e A select count(*) from dba_objects;
J o s select
select
count(*)
count(*)
from
from
dba_objects;
dba_objects;
select count(*) from dba_objects;
set termout on
exit;

9) Go back to the Top Consumers page in your Enterprise Manager session. Wait until
you see the TRACESERV service in the Active Services table, and enable tracing for
that service.
a) When you see TRACESERV in the Active Services table on the Top Services
page, select it, and click the Enable SQL Trace button.
b) On the Enable SQL Trace page, make sure Waits is set to TRUE, and Binds is set
to FALSE. Click OK.
c) Back to the Top Services page, you should see a confirmation message near the
top of the Top Consumers page.
10) When tracing for TRACESERV is enabled, execute the run_tracep1.sh script
from your terminal session. Observe your screen.
n s e
ice
el
[oracle@edrsr33p1-orcl Application_Tracing]$ ./run_tracep1.sh
b l
SQL*Plus: Release 11.1.0.6.0 - Production on Fri Apr 4 20:29:48
f e ra2008
t r a ns
o n -
s an
eฺ
a d11.1.0.6.0
Connected to:
Oracle Database 11g Enterprise Edition h
s ) Release
u i -
Production
m i rฺe Security,
n t G OLAP, Data Mining
ฺ
With the Partitioning, Oracle Label
gp options e
and Real Application Testing
d S tud
SQL>
b i a@ this
i a ( a use
SQL> connect jtrace/trace@TRACESERV
to
Connected.
SQL> Ab
o
ni alter session set workarea_size_policy=manual;
SQL>
n t o
e A Session altered.
Jos SQL>
Session altered.
SQL>
SQL> alter session set hash_area_size=5000;
Session altered.
SQL>
SQL>
Session altered.
SQL>
SQL>
SQL> set timing on

SQL>
SQL> select /*+ ORDERED USE_HASH(s2) */ count(*) from sales s1,
sales s2 where s1.cust_id=s2.cust_id;
COUNT(*)
----------
172878975
Elapsed: 00:01:19.25
SQL>
SQL>
SQL>
Connected.
SQL>
SQL> alter session set tracefile_identifier='mytraceS1';
n s e
i ce
Session altered.
b l el
Elapsed: 00:00:00.00
fer a
SQL>
a n s
SQL> set timing on
o n -tr
an
SQL>
SQL> select /*+ ORDERED USE_HASH(s2) S1 */ count(*) from sales s1,
sales s2 where s1.cust_id=s2.cust_id;
h s
a deฺ
e )
s Gui
COUNT(*) r ฺ
i ent
----------
p ฺ m
172878975
d g Stud
b i
Elapsed: 00:00:40.19a@ this
SQL>
i a (ja use
b fromtoOracle Database 11g Enterprise Edition Release
SQL> exit;
A
i o
Disconnected
n
nto11.1.0.6.0 - Production
e A With the Partitioning, Oracle Label Security, OLAP, Data Mining
J o s and Real Application Testing options

--------------------------------------------------------------
#!/bin/bash
export
--------------------------------------------------------------

[oracle@edrsr33p1-orcl Application_Tracing]$ cat run_tracep1.sql
set echo on
alter session set workarea_size_policy=manual;

alter session set sort_area_size=50000;
alter session set hash_area_size=5000;
set timing on
n s e
i ce
select /*+ ORDERED USE_HASH(s2) */ count(*) from sales s1, sales s2
b l el
where s1.cust_id=s2.cust_id;
fer a
a n s
o n -tr
an
h s
a deฺ
alter session set tracefile_identifier='mytraceS1';
e )
s Gui
set timing on r ฺ
i ent
p ฺ m d*/ count(*) from sales s1, sales
d g
select /*+ ORDERED USE_HASH(s2)
S t
S1u
b i a@ this
s2 where s1.cust_id=s2.cust_id;
exit;
i a (ja use
A b to
11) o
ni the run_tracep2.sh script from your terminal session. Observe your
Execute
t o
e An screen.
Jos [oracle@edrsr33p1-orcl Application_Tracing]$ ./run_tracep2.sh
Connected to:
Production
SQL>
Connected.
SQL>
Session altered.

SQL>
SQL> set timing on
SQL>
SQL> declare
2 c number := dbms_sql.open_cursor;
3 oname varchar2(50);
4 ignore integer;
5 begin
6 for i in 1 .. 5000 loop
7 dbms_sql.parse(c,'select object_name from dba_objects where
object_id = '||i , dbms_sql.native); -- use literal
8 dbms_sql.define_column(c, 1, oname, 50);
9 ignore := dbms_sql.execute(c);
10 if dbms_sql.fetch_rows(c)>0 then
11 dbms_sql.column_value(c, 1, oname);
n s e
12 end if;
i ce
13 end loop;
b l el
14 dbms_sql.close_cursor(c);
15 end;
fer a
16 /
a n s
o n -tr
an
Elapsed: 00:00:49.36
h s
a deฺ
SQL>
e )
s Gui
SQL> r ฺ
i ent
SQL>
p ฺ m d
d g
S t u
Connected.
SQL> b i a@ this
i a (ja usetsetracefile_identifier='mytraceS2';
SQL> alter session
A baltered.to
ni o
Session
ntoElapsed:
e A 00:00:00.00
J o s SQL>
SQL> declare
3 oname varchar2(50);
4 ignore integer;
5 begin
6 dbms_sql.parse(c,'select object_name from dba_objects where
object_id = :y' , dbms_sql.native); -- use bind var
7 for i in 1 .. 5000 loop
8 dbms_sql.bind_variable(c,':y',i);
9 dbms_sql.define_column(c, 1, oname, 50);
10 ignore := dbms_sql.execute(c);
11 if dbms_sql.fetch_rows(c)>0 then
12 dbms_sql.column_value(c, 1, oname);
13 end if;
14 end loop;
15 dbms_sql.close_cursor(c);
16 end;
17 /

Elapsed: 00:00:00.86
SQL>
SQL> exit;

--------------------------------------------------------------
#!/bin/bash
n s e
i ce
b l el
fer a
export
a n s
o n -tr
an

h s
a deฺ
e )
s Gui
r ฺ
i ent
--------------------------------------------------------------
p ฺ m
set echo on
d g Stud
b i a@ this
i a (jaset tracefile_identifier='mytraceP2';
u se
b
alter session
A to
o
ni timing on
ntoset
e A
J o s declare
c number := dbms_sql.open_cursor;
oname varchar2(50);
ignore integer;
begin
for i in 1 .. 5000 loop
dbms_sql.parse(c,'select object_name from dba_objects where
object_id = '||i , dbms_sql.native); -- use literal
dbms_sql.define_column(c, 1, oname, 50);
ignore := dbms_sql.execute(c);
if dbms_sql.fetch_rows(c)>0 then
dbms_sql.column_value(c, 1, oname);
end if;
end loop;
dbms_sql.close_cursor(c);
end;
/

declare
c number := dbms_sql.open_cursor;
oname varchar2(50);
ignore integer;
begin
dbms_sql.parse(c,'select object_name from dba_objects where
object_id = :y' , dbms_sql.native); -- use bind var
for i in 1 .. 5000 loop
dbms_sql.bind_variable(c,':y',i);
dbms_sql.define_column(c, 1, oname, 50);
dbms_sql.column_value(c, 1, oname);
n s e
end if;
i ce
end loop;
b l el
dbms_sql.close_cursor(c);
end;
fe r a
/
a n s
o n -tr
an
exit;
h s ฺ Observe your
aterminaldesession.
12) Execute the run_tracep3.sh script from )
your i
screen. i r ฺes nt Gu
g p ฺm ude
d t
S
./run_tracep3.sh
i @ his- Production on Fri Apr 4 20:32:53 2008

a11.1.0.6.0
SQL*Plus: Release
j a b e t
(
ia(c) 1982, s
u2007, Oracle. All rights reserved.
b
Copyright
t o
n i oA
to
e An Connected to:
Jos
Production
SQL>
Connected.
SQL>
Session altered.
SQL>
SQL> update sales set amount_sold=20000 where prod_id=13 and
cust_id=987;
2 rows updated.
SQL>
SQL> commit;

Commit complete.
SQL>
SQL>
Connected.
SQL>
SQL> create index sales_prod_cust_indx on sales(prod_id,cust_id);
Index created.
SQL>
Connected.
SQL>
n s e
i ce
b l el
Session altered.
fer a
SQL>
a n s
-tr
SQL> update sales set amount_sold=30000 where prod_id=13 and
n
o
an
cust_id=987;
2 rows updated.
h s
a deฺ
e )
s Gui
r ฺ
mi dent
SQL>
SQL> commit;
p ฺ
g Stu
d
Commit complete.
b i a@ this
SQL> a se
(jtrace/trace
b i
SQL> connecta to u
i o A
Connected.
n
toSQL>
A n SQL> drop index sales_prod_cust_indx;
o s e
J Index dropped.
SQL>
SQL>
SQL> exit;
--------------------------------------------------------------
#!/bin/bash

export

--------------------------------------------------------------
set echo on
update sales set amount_sold=20000 where prod_id=13 and cust_id=987;

n s e
i ce
commit;
b l el
fer a
connect trace/trace
a n s
o n -tr
an
create index sales_prod_cust_indx on sales(prod_id,cust_id);
h s
a deฺ
e )
s Gui
r ฺ
i ent
p ฺ m
d g Studwhere prod_id=13 and cust_id=987;
update sales set amount_sold=30000
commit; b i a@ this
ja e
ia ( to us
connect btrace/trace
i A
o index
to n
drop sales_prod_cust_indx;
A n
o s e
J exit;
13) Execute the run_tracep4.sh script from your terminal session. Observe your
screen.
Connected to:
Production
SQL>

Connected.
SQL>
Session altered.
SQL>
SQL> set timing on
SQL>
SQL> DECLARE
2 TYPE SalesCurTyp IS REF CURSOR;
3 v_sales_cursor SalesCurTyp;
4 sales_record sh.sales%ROWTYPE;
5 v_stmt_str VARCHAR2(200);
6 BEGIN
n s e
7 -- Dynamic SQL statement with placeholder:
i ce
8 v_stmt_str := 'select * from sh.sales where amount_sold>0';
b l el
9
10 -- Open cursor and specify bind argument in USING clause:
fer a
11 OPEN v_sales_cursor FOR v_stmt_str;
a n s
12
n -tr
o
an
13 -- Fetch rows from result set one at a time:
14 LOOP
15
h s
a deฺ
FETCH v_sales_cursor INTO sales_record;
16 )
s Gui
EXIT WHEN v_sales_cursor%NOTFOUND;
e
17 END LOOP; r ฺ
i ent
18
-- Close cursor: g p ฺ m d
19
d S t u
20
21 END; b i a@ this
CLOSE v_sales_cursor;
22 /
i a (ja use
A b o
proceduret successfully completed.
i o
PL/SQL
n
t o
n Elapsed: 00:00:26.84
e A
Jos
SQL>
SQL>
Connected.
SQL>
Session altered.
Elapsed: 00:00:00.00
SQL>
SQL> set timing on
SQL>
SQL> DECLARE
2 TYPE SalesCurTyp IS REF CURSOR;
3 TYPE SalesList IS TABLE OF sh.sales%ROWTYPE;
4 v_sales_cursor SalesCurTyp;
5 sales_List SalesList;
6 v_stmt_str VARCHAR2(200);
7 BEGIN
8 -- Dynamic SQL statement with placeholder:

9 v_stmt_str := 'select /* S4 */ * from sh.sales where
amount_sold>0';
10
11 -- Open cursor:
12 OPEN v_sales_cursor FOR v_stmt_str;
13
14 -- Fetch rows from result set one at a time:

15 LOOP
16 FETCH v_sales_cursor BULK COLLECT INTO Sales_List LIMIT
10000;
17 EXIT WHEN v_sales_cursor%NOTFOUND;
18 END LOOP;
19
20 -- Close cursor:
21 CLOSE v_sales_cursor;
22 END;
n s e
23 /
i ce
b l el
fer a
Elapsed: 00:00:02.09
a n s
SQL>
o n -tr
an
SQL>
SQL> exit;
h s
a deฺ
e )
s Gui
r ฺ
mi dent
p ฺ
g Stu
d
b i a@ this
--------------------------------------------------------------
i a (ja use
b
#!/bin/bash
A to
o
ni /home/oracle/solutions/Application_Tracing
ntocd
e A
export
--------------------------------------------------------------
set echo on
set timing on
DECLARE
TYPE SalesCurTyp IS REF CURSOR;

v_sales_cursor SalesCurTyp;
sales_record sh.sales%ROWTYPE;
v_stmt_str VARCHAR2(200);
BEGIN
-- Dynamic SQL statement with placeholder:
v_stmt_str := 'select * from sh.sales where amount_sold>0';
-- Open cursor and specify bind argument in USING clause:

OPEN v_sales_cursor FOR v_stmt_str;
-- Fetch rows from result set one at a time:

LOOP
FETCH v_sales_cursor INTO sales_record;
END LOOP;
n s e
-- Close cursor:
i ce
b l el
END;
/
fer a
a n s
n -tr
o
an
h s
a deฺ
e )
s Gui
set timing on r ฺ
i ent
p ฺ m d
DECLARE
TYPE SalesCurTyp @ d g S t u
a IS
TYPE SalesListbiIS TABLEth issh.sales%ROWTYPE;
REF CURSOR;
OF
(ja uSalesCurTyp;
v_sales_cursor
i a se
b
sales_List
A toVARCHAR2(200);
SalesList;
o
ni-- Dynamic SQL statement with placeholder:
v_stmt_str
ntoBEGIN
e A
J o s v_stmt_str := 'select /* S4 */ * from sh.sales where
amount_sold>0';
-- Open cursor:
OPEN v_sales_cursor FOR v_stmt_str;
-- Fetch rows from result set one at a time:

LOOP
FETCH v_sales_cursor BULK COLLECT INTO Sales_List LIMIT 10000;
END LOOP;
-- Close cursor:
END;
/
exit;

screen.

Connected to:
Production
n s e
SQL>
i ce
b l el
Connected.
fer a
SQL>
a n s
o n -tr
an
Session altered.
SQL>
h s
a deฺ
e
SQL> insert into sales2 select * from )
s Gui union all select *
sh.sales
r ฺ
mi dent
from sales;
1837686 rows created. dg p ฺ

S tu
b i a@ this
(ja use
SQL> commit;
b i a to
A
Commit complete.
o
ni
toSQL>
A n SQL>
o s e SQL> connect trace/trace@TRACESERV

J Connected.
SQL>
Session altered.
SQL>
SQL> insert into sales3 select * from sh.sales union all select *
from sales;
SQL> commit;
Commit complete.
SQL>
SQL> exit;

--------------------------------------------------------------
#!/bin/bash
export
n s e
i ce
b l el
fer a
--------------------------------------------------------------
a n s
o n -tr
an
set echo on
h s
a deฺ
e )
s Gui
r ฺ
i ent
p ฺ m d
d
insert into sales2 selectg * from u
S t sh.sales union all select * from
sales;
commit; b i a@ this
i a (ja use
A to
btrace/trace@TRACESERV
o
ni session set tracefile_identifier='mytraceS5';
connect
ntoalter
e A
J o s insert into sales3 select * from sh.sales union all select * from
sales;
commit;
exit;
screen.

Connected to:
Production
Connected to:
Production
SQL>
n s e
i ce
SQL>
b l el
Connected.
fer a
SQL>
a n s
SQL> update sales set amount_sold=amount_sold+1;
n -tr
o
an
Connected.
SQL>
h s
a deฺ
e )
s Gui
Session altered. r ฺ
i ent
p ฺ m d
SQL>
d g S t u
b i a@ this
918843 rows (
a ja use
updated.
b i to
SQL> A
o
ni exec dbms_lock.sleep(60);
ntoSQL>
e A
SQL>
SQL>
SQL> rollback;
Rollback complete.
SQL>
SQL> exit;

--------------------------------------------------------------
#!/bin/bash
export
n s e
i ce
b l el
sqlplus / as sysdba @run_tracep6a.sql &
fer a
sqlplus / as sysdba @run_tracep6b.sql
a n s
n -tr
o
an
--------------------------------------------------------------
h s
a deฺ
set echo on
e )
s Gui
r ฺ
connect trace/trace
p ฺ mi dent
d g Stu
b i a@ this
update sales set amount_sold=amount_sold+1;
i a (ja use
A b to
i o
rollback;
n
t o
n exit;
e A
Jos --------------------------------------------------------------
set echo on
set termout off
select cust_id, sum(amount_sold) from sales group by cust_id order

by cust_id;
set tournout on
exit;

screen.

Connected to:
Production
n s e
SQL>
i ce
b l el
Connected.
fer a
SQL>
a n s
o n -tr
an
Session altered.
SQL>
h s
a deฺ
SQL> declare
e )
s Gui
r ฺ
mi dent
3 custid number;
amount number;dg p ฺ
4
S tu
5
a@ this
ignore integer;
b i
(ja use
6 begin
7
b i a
dbms_sql.parse(c,'select
o
t -- use bind var
cust_id, sum(amount_sold) from
i o A
sales where cust_id=2
dbms_sql.native);
group by cust_id order by cust_id' ,
nton 8 dbms_sql.define_column(c, 1, custid);

e A 9 dbms_sql.define_column(c, 2, amount);
J o s 10
11
12 dbms_sql.column_value(c, 1, custid);
13 dbms_sql.column_value(c, 2, amount);
14 end if;
15 end;
16 /
SQL>
Connected.
SQL>
SQL> create index sales_cust_indx on sales(cust_id);
Index created.
SQL>
SQL> exit;

--------------------------------------------------------------
#!/bin/bash
n s e
export
i ce
b l el
fer a
a n s
n - tr
o
an
--------------------------------------------------------------
set echo on
h s
a deฺ
connect trace/trace@TRACESERV rฺe
)
s Gui
ฺ m i ent
d p
g Stud
declare b i a@ this
c number (
a jadbms_sql.open_cursor;
:=
u se
i to
b number;
custid number;
o A
ni ignore integer;
amount
ntobegin
e A
J o s dbms_sql.parse(c,'select cust_id, sum(amount_sold) from sales
where cust_id=2 group by cust_id order by cust_id' ,
dbms_sql.native); -- use bind var
dbms_sql.define_column(c, 1, custid);
dbms_sql.define_column(c, 2, amount);
dbms_sql.column_value(c, 1, custid);
dbms_sql.column_value(c, 2, amount);
end if;
end;
/
connect trace/trace
create index sales_cust_indx on sales(cust_id);
exit;
17) Disable tracing for your database.

a) Go back to the Top Services page in your Enterprise Manager session.
b) Ensure that TRACESERV is selected, and click the Disable SQL trace button.
c) Back to the Top Services page, you should see a confirmation message near the
top of the Top Consumers page.
18) Try to find out all the trace files that were generated to handle the previous seven
cases.
[oracle@edrsr33p1-orcl Application_Tracing]$ ./show_mytraces.sh
orcl_ora_19237_mytraceP0.trc
orcl_ora_19237_mytraceP0.trm
orcl_ora_19355_mytraceS1.trc
n s e
ce
orcl_ora_19355_mytraceS1.trm
eli
a b l
fe r
a n s
n - tr
o
s an
) h a deฺ
r ฺ e s Gui
p ฺ mi dent
g Stu
d
a@ this
b i
i a (ja use
b to
A
o n io
nt orcl_ora_19634_mytraceP7.trc
e A orcl_ora_19634_mytraceP7.trm
J o s [oracle@edrsr33p1-orcl Application_Tracing]$
--------------------------------------------------------------
#!/bin/bash
ls /u01/app/oracle/diag/rdbms/orcl/orcl/trace | grep mytrace
19) After you identify the location of those trace files, merge their content into one file
called mytrace.trc located in your
[oracle@edrsr33p1-orcl Application_Tracing]$ ./merge_traces.sh
--------------------------------------------------------------

#!/bin/bash
export
trcsess output=mytrace.trc service=TRACESERV

/u01/app/oracle/diag/rdbms/orcl/orcl/trace/*.trc
n s e
20) Use tkprof over the mytrace.trc file to generate a compiled trace output called
i ce
myreport.txt located into your
b l el
f er a
an s
- t r
[oracle@edrsr33p1-orcl Application_Tracing]$ ./tkprof_traces.sh
TKPROF: Release 11.1.0.6.0 - Production on Fri n on4 20:37:27 2008

Apr
s a
Copyright (c) 1982, 2007, Oracle. All h
) a
i d eฺ
rights reserved.
i r ฺes nt Gu
p ฺm ude
g
[oracle@edrsr33p1-orcl d
S t
Application_Tracing]$
@ s
j a b ia
e thi
--------------------------------------------------------------
b ia ( to us
oA
#!/bin/bash
n i
ntocd /home/oracle/solutions/Application_Tracing
e A
export
tkprof mytrace.trc myreport.txt
21) In addition, run TKPROF over the trace file that was generated for case 7 (step 16)
with the EXPLAIN option set to your TRACE account.
[oracle@edrsr33p1-orcl Application_Tracing]$ tkprof
/u01/app/oracle/diag/rdbms/orcl/orcl/trace/*mytraceP7.trc
myreport2.txt explain=trace/trace
TKPROF: Release 11.1.0.6.0 - Production on Fri Apr 4 20:40:22 2008

22) After this is done, interpret the trace generated for case 1 (step 10). What do you
observe, and what are your conclusions?

a) This case tries to illustrate the consequences of using manually sized SQL area
parameters, such as HASH_AREA_SIZE. The first statement was executed using
a very small HASH_AREA_SIZE value. The immediate consequence was the
number of temporary segments needed to execute the statement. Later, the same
statement was executed, but this time using the default SQL area parameters,
which were sized automatically by the system to handle the needs. You can see a
high disk value as well as a high number of waits for temporary segments for the
n s e
first execution, compared to the second one. i ce
b l el
select /*+ ORDERED USE_HASH(s2) */ count(*)
fe r a
from
a n s
-tr
sales s1, sales s2 where s1.cust_id=s2.cust_id
o n
call count cpu elapsed
s a n query current
disk
rows
)
------- ------ -------- ----------s----------
eฺ
ha id----------
r ฺ e G u ----------
mi0.00dent 0
----------
Parse 1 0.00
p
g Stuฺ 2 0
0 d
Execute 1
b i [email protected]
t h is 0.00 0 0 0
(j2a u78.23
se
0
Fetch
b i a o
79.16 806500 8874 0
1
i o A ------ t-------- ---------- ---------- ---------- ----------
-------
n
Anto----------
total 4 78.23 79.16 806500 8876 0
e
Jos
1
Misses in library cache during parse: 1

Optimizer mode: ALL_ROWS
Parsing user id: 182
Rows Row Source Operation

------- ---------------------------------------------------
1 SORT AGGREGATE (cr=8874 pr=806500 pw=806500 time=0 us)
172878975 HASH JOIN (cr=8874 pr=806500 pw=806500 time=1723284 us
cost=1392520 size=5737160286 card=220660011)
918843 TABLE ACCESS FULL SALES (cr=4437 pr=4433 pw=4433
time=7413 us cost=1363 size=14132716 card=1087132)
Elapsed times include waiting on following events:

Event waited on Times Max. Wait
Total Waited

---------------------------------------- Waited ---------- ---
---------
SQL*Net message to client 2 0.00
0.00
reliable message 1 0.00
0.00
enq: KO - fast object checkpoint 1 0.00

0.00
direct path read 86 0.00
0.00
direct path write temp 2524 0.00
0.00
direct path read temp 797634 0.00
1.70
SQL*Net message from client 2 0.00
0.00
n s e
********************************************************************
i ce
************
b l el
fer a
a n s
o n -tr
s an
h a deฺ
select /*+ ORDERED USE_HASH(s2) S1 */ count(*)
)
from
r ฺ e s Gui
mi dent
sales s1, sales s2 where s1.cust_id=s2.cust_id
p ฺ
g elapsed
d S tu
call
rows
count
b i a@ this
cpu disk query current
i a (ja --------
------- ------
u se ---------- ---------- ---------- ----------
b 1 to 0.00
----------
Parse A
0ni
o 0.00 0 2 0
ntoExecute
e A 1 0.00 0.00 0 0 0
J o s 0
Fetch 2 40.02 40.14 10028 8874 0
1
------- ------ -------- ---------- ---------- ---------- ----------
----------
total 4 40.03 40.14 10028 8876 0
1


------- ---------------------------------------------------
1 SORT AGGREGATE (cr=8874 pr=10028 pw=10028 time=0 us)
172878975 HASH JOIN (cr=8874 pr=10028 pw=10028 time=1430806 us
cost=6891 size=5737160286 card=220660011)


Total Waited
---------------------------------------- Waited ---------- ---
---------
0.00
direct path read 86 0.00
0.00
direct path write temp 166 0.00
0.00
direct path read temp 166 0.00
0.00
n s e
0.00
ice
********************************************************************
b l el
************
fe ra
a n s
23) Interpret the trace generated for case 2 (step 11). What do you observe, and what are
your conclusions?
o n -tr
a) For this case, the almost same statement was run 5000 a ntimes, but each time a
different literal was used to execute the query.a
h s eฺinefficient
This caused the system to parse
almost the same statement 5000 times, which )
s Gu is i d
extremely although the
r ฺ e
statement. However, another p ฺ mi dwaseinformation
time precision is too low to give accurate
statement
t
nalso about the parse time of each
d
time using a bind variable. S tu was parsed only 5000
gThe statement executed
once,
times, but this
and executed
i a @ i s
5000 times. This
j abbehavior
e thmuch more efficient than the previous one.
is
(
ia to u s
b
select object_name
from A
o n io
dba_objects where object_id = 1
nt
e A
J o s call count cpu elapsed disk query current
rows
------- ------ -------- ---------- ---------- ---------- ----------
----------
Parse 1 0.01 0.01 0 0 0
0
Execute 1 0.00 0.00 0 0 0
0
Fetch 1 0.00 0.00 0 3 0
0
------- ------ -------- ---------- ---------- ---------- ----------
----------
total 3 0.01 0.01 0 3 0
0

Parsing user id: 182 (recursive depth: 1)

------- ---------------------------------------------------
0 VIEW DBA_OBJECTS (cr=3 pr=0 pw=0 time=0 us cost=5 size=158
card=2)
0 UNION-ALL (cr=3 pr=0 pw=0 time=0 us)
0 FILTER (cr=3 pr=0 pw=0 time=0 us)
0 NESTED LOOPS (cr=3 pr=0 pw=0 time=0 us cost=5 size=109
card=1)
card=1)
0 TABLE ACCESS BY INDEX ROWID OBJ$ (cr=3 pr=0 pw=0
1 INDEX RANGE SCAN I_OBJ1 (cr=2 pr=0 pw=0 time=0 us
cost=2 size=0 card=1)(object id 36)
0 INDEX RANGE SCAN I_USER2 (cr=0 pr=0 pw=0 time=0 us
n s e
i ce
0 TABLE ACCESS BY INDEX ROWID IND$ (cr=0 pr=0 pw=0 time=0
b l el
us cost=2 size=8 card=1)
0
fe
INDEX UNIQUE SCAN I_IND1 (cr=0 pr=0 pw=0 time=0 us r a
a n s
0 -tr
NESTED LOOPS (cr=0 pr=0 pw=0 time=0 us cost=2 size=28
n
o
an
card=1)
0 INDEX FULL SCAN I_USER2 (cr=0 pr=0 pw=0 time=0 us
h s
a deฺ
0 )
s Gui
INDEX RANGE SCAN I_OBJ4 (cr=0 pr=0 pw=0 time=0 us
e
r ฺ
mi dent
0 ฺ
FILTER (cr=0 pr=0 pw=0 time=0 us)
p
g Stu
0
d
NESTED LOOPS (cr=0 pr=0 pw=0 time=0 us cost=1 size=83
card=1)
0 b i a@ this
INDEX FULL SCAN I_LINK1 (cr=0 pr=0 pw=0 time=0 us
i a (ja use
0b to
INDEX RANGE SCAN I_USER2 (cr=0 pr=0 pw=0 time=0 us
n i oA
nto********************************************************************
e A
J o s ************
select object_name
from
call count cpu elapsed disk query current

rows
------- ------ -------- ---------- ---------- ---------- ----------
----------
Parse 1 0.00 0.00 0 0 0
0
Execute 1 0.00 0.00 0 0 0
0
Fetch 1 0.00 0.00 0 5 0
1
------- ------ -------- ---------- ---------- ---------- ----------
----------

total 3 0.00 0.00 0 5 0
1

...
select object_name
from

n s e
rows
i ce
------- ------ -------- ---------- ---------- ---------- ----------
b l el
----------
Parse 1 0.00 0.00 0 0
fe r a 0
0
a n s
Execute 1 0.00 0.00 0
o n -tr 0 0
a0 n
0
Fetch 1 0.00 0.00 5 0
1
h a s e ฺ
------- ------
s )
-------- ---------- ----------
u i d---------- ----------
---------- r ฺ
i0.00e tG 0
total 3 0.00
p ฺ m d e n 5 0
1
d g S t u
@ s
Misses in library
j a b e thi parse: 1
iacache during
ia (id:to182us (recursive depth: 1)
Parsing buser
n i o A Row Source Operation

nto------- ---------------------------------------------------
Rows
e A
J o s 1 VIEW DBA_OBJECTS (cr=5 pr=0 pw=0 time=0 us cost=5 size=158
card=2)
card=1)
card=1)
0 INDEX UNIQUE SCAN I_IND1 (cr=0 pr=0 pw=0 time=0 us

card=1)

card=1)
0 INDEX FULL SCAN I_LINK1 (cr=0 pr=0 pw=0 time=0 us
********************************************************************
************
n s e
i ce
b l el
fer a
a n s
o n -tr
s an
select object_name
) h a deฺ
from
r ฺ e s Gui
mi dent
dba_objects where object_id = :y
p ฺ
g elapsed
d S tu
call
rows
count
b i a@ this
cpu disk query current
i a (ja --------
------- ------
u se ---------- ---------- ---------- ----------
b 1 to 0.00
----------
Parse A
0ni
o 0.00 0 0 0
ntoExecute 5000
e A 0.20 0.18 0 0 0
J o s 0
Fetch 5000 0.24 0.24 0 26837 0
4928
------- ------ -------- ---------- ---------- ---------- ----------
----------
total 10001 0.45 0.43 0 26837 0
4928

Misses in library cache during execute: 1

------- ---------------------------------------------------
0 VIEW DBA_OBJECTS (cr=3 pr=0 pw=0 time=0 us cost=5 size=158
card=2)
card=1)

card=1)

0 INDEX UNIQUE SCAN I_IND1 (cr=0 pr=0 pw=0 time=0 us
card=1)
n s e
ice
b l el
fe ra
a n s
o n -tr
an
card=1)
0 INDEX FULL SCAN I_LINK1 (cr=0 pr=0 pw=0 time=0 us
h s
a deฺ
0 )
s Gui
INDEX RANGE SCAN I_USER2 (cr=0 pr=0 pw=0 time=0 us
e
cost=1 size=4 card=1)(object id 47) r ฺ
i ent
p ฺ m d
d g t u
********************************************************************
S
************
b i a@ this
i a (jagenerated
24) Interpret the trace
u sefor case 3 (step 12). What do you observe, and what are
b
your conclusions?
A to
o
ni If you look closely at this case, you see that you access the complete table to
a)
t o
e An update only two rows. This is very inefficient and the alternate case is much better
Jos
as it uses an index to speed up the retrieval of the rows that need to be updated.
update sales set amount_sold=20000
where
prod_id=13 and cust_id=987

rows
------- ------ -------- ---------- ---------- ---------- ----------
----------
Parse 1 0.00 0.00 0 1 0
0
Execute 1 0.09 0.09 2442 4441 4
2
Fetch 0 0.00 0.00 0 0 0
0
------- ------ -------- ---------- ---------- ---------- ----------
----------
total 2 0.09 0.10 2442 4442 4
2


------- ---------------------------------------------------
0 UPDATE SALES (cr=4441 pr=2442 pw=2442 time=0 us)
2 TABLE ACCESS FULL SALES (cr=4441 pr=2442 pw=2442 time=17

Total Waited
---------------------------------------- Waited ---------- ---
n s e
---------
i ce
db file scattered read 63 0.00
b l el
0.01
db file sequential read 3 0.00
fer a
0.00
a n s
SQL*Net message to client 1
n -tr
0.00
o
an
0.00
0.00
h s
a deฺ
)
s Gui
********************************************************************
e
r ฺ
mi dent
************
p ฺ
g Stu
d
b i a@ this
a
(jset se
b i
update salesa to u
amount_sold=30000
where A
i o
nprod_id=13
n t o and cust_id=987
e A
Jos call
rows
count cpu elapsed disk query current
------- ------ -------- ---------- ---------- ---------- ----------

----------
Parse 1 0.00 0.00 0 2 0
0
Execute 1 0.00 0.00 0 3 3
2
Fetch 0 0.00 0.00 0 0 0
0
------- ------ -------- ---------- ---------- ---------- ----------
----------
total 2 0.00 0.00 0 5 3
2


------- ---------------------------------------------------
0 UPDATE SALES (cr=3 pr=0 pw=0 time=0 us)
2 INDEX RANGE SCAN SALES_PROD_CUST_INDX (cr=3 pr=0 pw=0
time=2 us cost=3 size=78 card=2)(object id 75372)

Total Waited
---------------------------------------- Waited ---------- ---
---------
0.00
0.00
********************************************************************
n s e
************
ice
b l
25) Interpret the trace generated for case 4 (step 13). What do you observe, and what areel
your conclusions? fe r a
a n s
o n -tr
a) In this case, the first statement does not use the fetching mechanism appropriately.
One fetch operation is done for every single row retrieved. This is also very
an
inefficient. The alternate case is doing 92 fetches to retrieve the same amount of
s
h a deฺ
rows. This technique is called array fetch.
)
r ฺ e s Gui
mi dent
select *
from
p ฺ
g Stu
d
sh.sales where amount_sold>0
b i a@ this
call
a (ja uscpu
count
i
e elapsed disk query current
rows b o
A ------ t-------- ---------- ---------- ---------- ----------
i o
-------
n
nto----------
e A Parse 1 0.02 0.02 0 0 0
J o s 0
Execute 1 0.00 0.00 0 0 0
0
Fetch 918844 5.89 5.93 0 918944 0
918843
------- ------ -------- ---------- ---------- ---------- ----------
----------
total 918846 5.92 5.95 0 918944 0
918843


------- ---------------------------------------------------
918843 PARTITION RANGE ALL PARTITION: 1 28 (cr=918944 pr=0 pw=0
918843 TABLE ACCESS FULL SALES PARTITION: 1 28 (cr=918944 pr=0
pw=0 time=27743 us cost=547 size=26646447 card=918843)

********************************************************************
************
select /* S4 */ *
from
sh.sales where amount_sold>0

rows
------- ------ -------- ---------- ---------- ---------- ----------
----------
n s e
Parse 1 0.00 0.00 0 0 0
i ce
0
b l e0 l
Execute
0
1 0.00 0.00 0 0
fe r a
Fetch 92 2.06 2.06 0 1811
a n s 0
918843
o n -tr ----------
a0 n 1811
------- ------ -------- ---------- ---------- ----------
----------
total 94 2.06 2.06
h a s e ฺ 0
918843
s) Gui d
r ฺ e
iparse:en1 t
Misses in library cache during
Optimizer mode: ALL_ROWS g p ฺ m d
Parsing user id: 182@ d S t u
b i a t h is
(recursive depth: 1)
Rows a
Row(jSource se
a to uOperation
------- bi---------------------------------------------------
n i o A PARTITION
918843 RANGE ALL PARTITION: 1 28 (cr=1811 pr=0 pw=0
nto 918843 TABLE ACCESS FULL SALES PARTITION: 1 28 (cr=1811 pr=0 pw=0
e A
Jos
********************************************************************
************
your conclusions?
a) Here, the first statement incurs too many recursive calls to allocate extents to the
table. This is because the tablespace in which it is stored is not correctly set up for
extent allocations. The alternate case is much better as you can see the number of
disk values going way down compared to the first case. Also the number of
recursive statements in the second case should be much lower than the first one.
insert into sales2 select * from sh.sales union all select * from
sales

rows

------- ------ -------- ---------- ---------- ---------- ----------
----------
Parse 1 0.00 0.01 0 2 0
0
Execute 1 5.02 5.41 4432 28134 148866
1837686
Fetch 0 0.00 0.00 0 0 0

0
------- ------ -------- ---------- ---------- ---------- ----------
----------
total 2 5.02 5.42 4432 28136 148866
1837686

n s e
i ce
b l el
------- ---------------------------------------------------
fe
0 LOAD TABLE CONVENTIONAL (cr=44323 pr=4433 pw=4433 time=0r a
us)
a n s
1837686 -tr
UNION-ALL (cr=6160 pr=4432 pw=4432 time=58107 us)
o n
an
918843
h s
a deฺ
TABLE ACCESS FULL SALES PARTITION: 1 28 (cr=1719 pr=0
)
s Gui
e
r ฺ
mi dent
p
g Stuฺ
d
b i
Elapsed times includea@waiting
t h is on following events:
Event waited
i a (jaon use Times Max. Wait
b
Total Waited
A to
n i o
---------------------------------------- Waited ---------- ---
nto db file scattered read

---------
e A 51 0.00
J o s 0.03
log file switch completion 2 0.00
0.00
log buffer space 8 0.09
0.23
0.00
0.00
********************************************************************
************
select file#
from
file$ where ts#=:1

rows
------- ------ -------- ---------- ---------- ---------- ----------
----------

Parse 1812 0.02 0.02 0 0 0
0
Execute 1812 0.02 0.03 0 0 0
0
Fetch 3624 0.05 0.05 0 7242 0
1812
------- ------ -------- ---------- ---------- ---------- ----------

----------
total 7248 0.11 0.10 0 7242 0
1812

Misses in library cache during execute: 1
Optimizer mode: CHOOSE
Parsing user id: SYS (recursive depth: 1)
n s e
i ce
------- ---------------------------------------------------
b l el
1 TABLE ACCESS FULL FILE$ (cr=4 pr=0 pw=0 time=0 us cost=2
size=6 card=1)
fer a
a n s
- tr
********************************************************************
n
o
an
************
h s
a deฺ
e )
s Gui
r ฺ
p ฺ mi dent
d g Stu
b i a@ this
i a (ja use
Ab to
o n io
insert into sales3 select * from sh.sales union all select * from
sales
nt
e A
J o s call count cpu elapsed disk query current
rows
------- ------ -------- ---------- ---------- ---------- ----------
----------
Parse 1 0.00 0.00 0 1 0
0
Execute 1 3.64 5.13 1087 22471 77748
1837686
Fetch 0 0.00 0.00 0 0 0
0
------- ------ -------- ---------- ---------- ---------- ----------
----------
total 2 3.65 5.14 1087 22472 77748
1837686


------- ---------------------------------------------------
0 LOAD TABLE CONVENTIONAL (cr=22544 pr=1087 pw=1087 time=0
us)
918843 TABLE ACCESS FULL SALES PARTITION: 1 28 (cr=1719 pr=0


Total Waited
---------------------------------------- Waited ---------- ---
n s e
---------
ice
buffer busy waits 3 0.00
b l el
0.00
undo segment extension 1 0.00
fe ra
0.00
a n s
log file switch completion 5
n - 0.08 tr
o
an
0.12
log buffer space 30 0.15
1.24
h s
a deฺ
db file scattered read
e )
s Gui 25 0.00
r ฺ
mi dent
0.00
p ฺ
log file switch (private strand flush incomplete)
g Stu
d 313 0.05
0.10
b i a@ this
0.00
i a (ja use
Ab to
o n io
0.00
********************************************************************
nt
e A ************
J o s 27) Interpret the trace generated for case 6 (step 15). What do you observe, and what are
your conclusions?
a) This case is more difficult to understand. Here, you select a table that is entirely
locked by another transaction. This forces the query to generate consistent read
blocks for almost the entire table causing undo segments to be accessed. This is
materialized by an important disk value, and almost no current blocks.
select cust_id, sum(amount_sold)
from
sales group by cust_id order by cust_id

rows
------- ------ -------- ---------- ---------- ---------- ----------
----------
Parse 1 0.00 0.00 0 1 0
0

Execute 1 0.00 0.00 0 0 0
0
Fetch 472 2.38 2.39 5333 978282 0
7059
------- ------ -------- ---------- ---------- ---------- ----------
----------
total 474 2.39 2.39 5333 978283 0

7059


------- ---------------------------------------------------
7059 SORT GROUP BY (cr=978282 pr=5333 pw=5333 time=64 us
n s e
cost=1422 size=28265432 card=1087132)
i ce
b l el
fe r a
a n s
n - tr
o
an
Total Waited
----------------------------------------
h s
a deฺ
Waited ---------- ---
---------
e )
s Gui
r ฺ
mi dent
0.00
p ฺ
g Stu
d
db file sequential read 3827 0.00
0.03
b i a@ this
db file scattered read 37 0.00
0.01
i a (ja use
Ab to
latch: cache buffers lru chain 1 0.00
o io
0.00
nSQL*Net message from client 472 0.00
nt
e A 0.02
J o s ********************************************************************
************
your conclusions?
a) For case 7, you should compare the content in myreport.txt with the content
in myreport2.txt. You should see that an index was added after the first trace
was generated. This situation can cause confusion especially if the trace does not
contain an execution plan to begin with. This is what you can see from within
myreport.txt:
select cust_id, sum(amount_sold)
from
sales where cust_id=2 group by cust_id order by cust_id

rows

------- ------ -------- ---------- ---------- ---------- ----------
----------
Parse 1 0.00 0.00 0 1 0
0
Execute 1 0.00 0.00 0 0 0
0
Fetch 1 0.05 0.05 0 4441 0

1
------- ------ -------- ---------- ---------- ---------- ----------
----------
total 3 0.06 0.06 0 4442 0
1

n s e
i ce
b l el
------- ---------------------------------------------------
fer
1 SORT GROUP BY NOSORT (cr=4441 pr=0 pw=0 time=0 us cost=1366a
size=1872 card=72)
a n s
176 -tr
TABLE ACCESS FULL SALES (cr=4441 pr=0 pw=0 time=153 us
o n
an
cost=1366 size=1872 card=72)
h s
a deฺ
e )
s Gui
r ฺ
i ent
ฺ m
********************************************************************
p
************
d g Stud
i
b) This is what youbseea@ h is
from myreport2.txt:
t
a (ja use
SQL ID : i51xxq509gnbbc
selectA b
cust_id, o
tsum(amount_sold)
i o
tonsales where cust_id=2 group by cust_id order by cust_id
from
A n
o s e
J call count cpu elapsed disk query current
rows
------- ------ -------- ---------- ---------- ---------- ----------
----------
Parse 1 0.00 0.00 0 1 0
0
Execute 2 0.00 0.00 0 0 0
0
Fetch 1 0.05 0.05 0 4441 0
1
------- ------ -------- ---------- ---------- ---------- ----------
----------
total 4 0.06 0.06 0 4442 0
1

Parsing user id: 182 (TRACE) (recursive depth: 1)

------- ---------------------------------------------------
1 SORT GROUP BY NOSORT (cr=4441 pr=0 pw=0 time=0 us cost=1366
size=1872 card=72)
176 TABLE ACCESS FULL SALES (cr=4441 pr=0 pw=0 time=153 us
cost=1366 size=1872 card=72)
Rows Execution Plan

------- ---------------------------------------------------
0 SELECT STATEMENT MODE: ALL_ROWS
1 SORT (GROUP BY NOSORT)
176 TABLE ACCESS (BY INDEX ROWID) OF 'SALES' (TABLE)
0 INDEX MODE: ANALYZED (RANGE SCAN) OF 'SALES_CUST_INDX'
(INDEX)
n s e
i ce
b l el
Event waited on
Total Waited
Times Max. Wait
fer a
---------------------------------------- n s
Waited ---------- ---
a
---------
o n -tr
an
0.00
SQL*Net message from client
h s
a deฺ
1 0.00
0.00
e )
s Gui
r ฺ
mi dent
********************************************************************
************
p ฺ
g Stuby executing the at_cleanup.sh script
29) You can now clean up your d
environment
from your terminal b i a@ this
window.
i a (ja useApplication_Tracing]$ ./at_cleanup.sh
b to
A
ni o
n t oSQL*Plus: Release 11.1.0.6.0 - Production on Fri Apr 4 20:45:24 2008
e A Copyright (c) 1982, 2007, Oracle. All rights reserved.

Jos
Connected to:
Production
SQL>
SQL> drop user trace cascade;
User dropped.
SQL>
SQL> drop tablespace tracetbs including contents and datafiles;
Tablespace dropped.
SQL>
SQL> drop tablespace tracetbs3 including contents and datafiles;

Tablespace dropped.
SQL>
SQL> exit;
--------------------------------------------------------------
#!/bin/bash
n s e
i ce
b l el
fer a
a n s
export
o n -tr
an
h s
a deฺ
)
s Gui
rm /u01/app/oracle/diag/rdbms/orcl/orcl/trace/*mytrace*.*
e
r ฺ
rm mytrace.trc
p ฺ mi dent
d g Stu
rm myreport.txt
b i a@ this
i a (ja use
rm myreport2.txt
A b to
o
ni $ORACLE_HOME/network/admin/tnsnames.ora.bak1
rm $ORACLE_HOME/network/admin/tnsnames.ora
ntomv
e A
J o s $ORACLE_HOME/network/admin/tnsnames.ora
sqlplus / as sysdba @at_cleanup.sql
--------------------------------------------------------------
set echo on
drop user trace cascade;
drop tablespace tracetbs including contents and datafiles;
drop tablespace tracetbs3 including contents and datafiles;
exit;

n s e
i ce
b l el
fer a
a n s
n - tr
o
s an
) h a deฺ
r ฺ e s Gui
p ฺ mi dent
d g Stu
b i a@ this
i a (ja use
Ab to
o n io
nt
e A
J o s

Practice 11-1: Proactively Tuning High-Load SQL Statements
In this practice, you use the SQL Tuning Advisor to help you tune problematic SQL
statements.
1) Connect as SYSDBA through Database Control and navigate to the Performance tab
of the Database Control Home page. On the Performance tabbed page, make sure that
the View Data field is set to Real Time: 15 second Refresh. After this is done, open a
terminal emulator window connected as the oracle user. When this is done, change
your current directory to your lab directory: cd
$HOME/solutions/SQL_Tuning_Advisor. Then enter the following
command from the OS prompt: ./setup_dina.sh
[oracle@edrsr33p1-orcl SQL_Tuning_Advisor]$ ./setup_dina.sh
n s e
[oracle@edrsr33p1-orcl SQL_Tuning_Advisor]$ ./setup_dina.sh
ice
b l el
fe r a
a n s
Grant succeeded.
n - tr
o
s an
Session altered.
) h a deฺ
r ฺ e s Gui
ฺ i ent
mcompleted.
p
g Stud
d
a@ this completed.
PL/SQL procedurebisuccessfully
i a (ja use
A b to
i
Usero altered.
nton
e A
J o s User altered.
Index dropped.
drop index sales_time_idx

*
ERROR at line 1:
ORA-01418: specified index does not exist
Index created.
[oracle@edrsr33p1-orcl SQL_Tuning_Advisor]$
--------------------------------------------------------------
#!/bin/bash

(continued)
cd /home/oracle/solutions/SQL_Tuning_Advisor
export
sqlplus -s /NOLOG <<EOF
set echo on
connect / as sysdba
n s e
i ce
exec DBMS_WORKLOAD_REPOSITORY.CREATE_SNAPSHOT();
b l el
fer a
grant dba to SH;
a n s
n -tr
rem -- event to allow setting very short Flushing interval
o
s a n forever, level
alter session set events '13508 trace name context
1';
) h a deฺ
r ฺ e s Gui
p
rem -- change INTERVAL setting ฺ mito 2dminutes
e nt
g Sto
rem -- change RETENTIONdsetting tu 6 hours (total of 180 snapshots)
@ his
ia360);
execute dbms_workload_repository.modify_snapshot_settings(interval
=> 2,retentiona=>
j b e t
(
ia to u s
b
n i
rem o--Aplay with ADDM sensitiveness
ntoexec
e A dbms_advisor.set_default_task_parameter('ADDM','DB_ACTIVITY_MIN',30)
J o s ;
alter user sh account unlock;

alter user sh identified by sh;
connect sh/sh
drop index sales_time_bix;

drop index sales_time_idx;
create index sales_time_idx on sales(time_id) compute statistics;
EOF
2) After this is executed, execute the start_dinas.sh script by using the following
command: ./start_dinas.sh
This script starts the workload used for this lab.

(continued)
[oracle@edrsr33p1-orcl SQL_Tuning_Advisor]$ ./start_dinas.sh
Started stream with pid=30479

--------------------------------------------------------------
#!/bin/bash
n s e
i ce
el
a b l
fer
a n s
export
n -tr
o
an
s
) h a deฺ
STREAM_NUM=0
r ฺ e s Gui
MAX_STREAM=6
p ฺ mi dent
PIDLST=""
d g Stu
while [ $STREAM_NUM
b i a@ h is
-lt $MAX_STREAM
t ]; do
i
# one morea (ja use
let A b to
STREAM_NUM="STREAM_NUM+1"
o
nt oni# start one more stream
eA sqlplus -S sh/sh @dina.sql &
Jos # remember PID

PIDLST="$! $PIDLST"
echo "Started stream with pid=$!"
done
#
# Save PID List
#
echo $PIDLST > /tmp/dina_pids
--------------------------------------------------------------
DECLARE
n number;
BEGIN
select /*+ ORDERED USE_NL(c) FULL(c) FULL(s)*/ count(*) into n
from sales s, customers c

(continued)
where c.cust_id = s.cust_id and CUST_FIRST_NAME='Dina'
order by time_id;
end loop;
END;
/
3) When the start_dinas.sh script completes, observe the Performance tabbed

page for around 10 minutes (5 snapshot icons). What are your conclusions?
a) You should see that the workload activity goes up very quickly. Because the CPU
used by the workload is very close to the maximum CPU available on your
system, there must be an issue with this workload. Because the most important
area corresponding to a wait class is the Other wait class, the issue must be
n s e
associated to that class. Note that the snapshot interval is now around two ice
minutes.
b l el
4) Fix the problem. fe r a
a n s
o n -tr
a) The fastest way to determine the problem is by looking at an Automatic Database
Diagnostic Monitor (ADDM) report analysis executed during the problematic
an
period. Then, by following its analysis, ADDM should guide you through the
s
process of fixing the problem. ) h a deฺ
b) Using the Database Control Home ipage, r ฺesthere G ui different ways to identify
ฺm udent
are two
the correct ADDM analysisgtask: p t
d to theSproblematic
i @
i) If the time corresponding
a detected h i s time period corresponds with the
j
latest ADDM a b run e t by Database Control, you should find the link
( u s
A bia section
corresponding
Summary
toto the
of
correct performance analysis directly in the Diagnostic
the Database Control Home page. Note that you should
n i o wait around 8 to 10 minutes
Anto before the Diagnostic Summary section is
refreshed with the correct ADDM analysis. If you are in this case, click the
e
Jos link corresponding to the number of findings right next to the ADDM
Findings row. This takes you to the corresponding Automatic Database
Diagnostic Monitor (ADDM) page.
ii) If not, you should open the Advisor Central page and search for the correct
ADDM task. This is how you can retrieve the task from the Advisor Central
page:
(1) On the Database Control Home page, click the Advisor Central link.
(2) On the Advisor Central page, in the search section, select ADDM from
the Advisory Type drop-down list, and Last 24 Hours from the
Advisor Runs drop-down list.
(3) After this is done, click Go.
(4) Then select the ADDM task corresponding to the time of the
problematic period.

(continued)
(5) This takes you to the corresponding Automatic Database Diagnostic
Monitor (ADDM) page.
c) On the Automatic Database Diagnostic Monitor (ADDM) page, you should see
two main findings: CPU Usage and Top SQL by DB Time. Both should be close
to 100%. If they are not, ensure that what you see is the latest ADDM analysis.
d) Click the “Top SQL by DB Time” link.
e) On the “Performance Finding Details: Top SQL by DB Time” page, click Show
All Details link. You should see something similar to the following: Run SQL
Tuning Advisor on the SQL statement with SQL_ID "5mxdwvuf9j3vp". Next to
this recommendation, click the Run Advisor Now button.
f) Wait on the Processing: SQL Tuning Advisor Task SQL_TUNING_… page for a n s e
ice
while.
b l el
g) You are automatically directed to the Recommendations for SQL
fe r a
n
ID:5mxdwvuf9j3vp” page from where you should see two recommendations-one
a s
for a SQL Profile and one for an index creation.
o n -tr
h) You can investigate the consequence of implementing
s a nthe recommended profile
by comparing execution plans before and after
) h i eฺ
aprofiledimplementation. You can
do so by clicking the “Compare explain
SQL Profile row. Because the potential i r ฺesbenefit
n t of
u icon for the corresponding
plan” eyeglass
G using the proposed SQL profile
p ฺ m d e
g Costs),
is very high (see both computed
d S tu you implement the SQL profile.
i
i) On the Compare Explain
b a@ Planst h ispage, click the Recommendations for SQL
(ja ulocator
ID:5mxdwvuf9j3vp
i a se link.
j) Back A to
bto the Recommendations for SQL ID:5mxdwvuf9j3vp page, ensure that the
n i o
SQL Profile row is selected, and click Implement.
n t o
e A k) On Confirmation page, click Yes.
Jos l) On Recommendations for SQL ID:5mxdwvuf9j3vp page, you should see the
following message at its top: “Confirmation: The recommended SQL Profile has
been created successfully”.
m) Click Return.
n) On SQL Tuning Results:TASK, click the Database Instance: orcl locator link.
o) Back to the Database Home Page, click the Performance tab.
5) After following the SQL Tuning Advisor recommendation to implement a SQL
Profile, how can you quickly verify that the problem is solved?
a) On the Performance tabbed page, you expect to see a dramatic drop for the Other
wait class in the Average Active Sessions graph. However, when you view the
graph you see that not has changed.
6) How do you interpret the result you see, and how would you ensure that the new
profile is taken into account?

(continued)
a) A profile is taken into account the next time you execute the corresponding
statement. Because the SQL statement for which you created a profile takes a long
time to execute, you should wait for a long time to see the benefit. To quickly see
the benefit, stop and restart your workload. This executes the same SQL
statements again, and the profile should be used automatically this time.
[oracle@edrsr33p1-orcl SQL_Tuning_Advisor]$ ./stop_dinas.sh
Killing stream with pid=30486
DECLARE
*
ERROR at line 1:
ORA-00028: your session has been killed
n s e
ice
DECLARE
b l el
*
fe ra
ERROR at line 1:
a n s
n - tr
o
s an
DECLARE
) h a deฺ
*
r ฺ e s Gui
mi killed nt
ERROR at line 1:
ORA-00028: your session has ฺbeen e
d p
g Stu d
b i a@ t h is
(ja use
DECLARE
b i a to
*
o A
ni
ERROR at line 1:
ntoORA-00028: your session has been killed
e A
J o s Killing stream with pid=30480
DECLARE
*
ERROR at line 1:

DECLARE
*
ERROR at line 1:
[oracle@edrsr33p1-orcl SQL_Tuning_Advisor]$ ./start_dinas.sh


(continued)
--------------------------------------------------------------
#!/bin/bash
n s e
i ce
el
a b l
export
fer
a n s
-tr
o n
PIDLST=`cat /tmp/dina_pids`
s an
) h a deฺ
#
r ฺ e s Gui
# Kill all these processes
# p ฺ mi dent
for PID in $PIDLST; do dg tu
S
b ia@@kill_dina.sql
echo "Killing stream
sqlplus / asasysdba
withis
t h pid=$PID"
sqlplus a ( u s e
j @/tmp/drop_dina.sql $PID >> /tmp/stop_dina.log 2>&1
b i $PID >>to/tmp/stop_dina.log >>

/nolog /tmp/stop_dina.log 2>&1
oA
kill -9 2>&1
n i
done
nto
e A --------------------------------------------------------------
J o s set head off

set timing off
set feedback off;
set pagesize 0
set verify off
spool /tmp/drop_dina.sql;
select 'connect / as sysdba;' from dual;
select 'alter system kill session ''' || sid || ',' || serial# ||

''';'
from v$session
where process=&1;
select 'exit;' from dual;
spool off
exit;

(continued)
b) Go back to the Performance tabbed page.
c) On the Performance page, you should now see the benefit of the SQL Profile.
There is no longer any Other waits in the Average Active Sessions graph.
7) How would you make sure that the SQL Profile was implemented?
a) On the Average Active Sessions graph, click the Other wait category in the
caption.
b) On the Active Sessions Waiting: Other page, you should see that your statement
waits for the latch free event. You can see that from the ACTIVITY(%) column.
c) Click the 5mxdwvuf9j3vp SQL Id link in the Top SQL table.
d) On the SQL Details: 5mxdwvuf9j3vp page, click the Plan Control tab in the n s e
i ce
Details section.
b l el
r
e) You should see that this statement is associated to a SQL Profile that was
fe a
n s
manually implemented. It is part of the DEFAULT category and is ENABLED.
a
n -tr
8) Clean up your environment by executing the following commands from your
o
command-line window:
s an
./stop_dians.sh
) h a deฺ
./cleanup_dina.sh
r ฺ e s Gui
p ฺ mi dent ./stop_dinas.sh
d g Stu
DECLARE
b i a@ this
(ja1: use
*
b
ORA-00028:i a
ERROR at line
your o
tsession has been killed
A
io stream with pid=31735
t o n
Killing
A n DECLARE
o s e *
J ERROR at line 1:

DECLARE
*
ERROR at line 1:

DECLARE
*
ERROR at line 1:

DECLARE
*
ERROR at line 1:

(continued)
DECLARE
*
ERROR at line 1:
[oracle@edrsr33p1-orcl SQL_Tuning_Advisor]$ ./cleanup_dina.sh
Revoke succeeded.
n s e
i ce
el
a b l
fer
a n s
-tr
o n
s an
specify directory path for the data files
) h i d eฺ 5:
a as parameter
i r es files
ฺlog t G u
ฺm uden
writeable directory path for the as parameter 6:
g p
d 7:St
specify version as parameter
i a @ h is
b t
i a (ja use
Session altered.
A b to
o
ni dropped.
ntoUser
e A
J o s old
new
1: CREATE USER sh IDENTIFIED BY &pass
1: CREATE USER sh IDENTIFIED BY sh
User created.

User altered.

User altered.
Grant succeeded.

(continued)
Grant succeeded.
...
Grant succeeded.
Grant succeeded.
Connected.
n s e
i ce
Grant succeeded.
b l el
fer a
old
s
1: CREATE OR REPLACE DIRECTORY data_file_dir AS '&data_dir'
a n
-tr
n
o
Directory created.
s an
) h a deฺ
old 1: CREATE OR REPLACE DIRECTORY
r ฺ e s Gui AS '&log_dir'
log_file_dir
new i elog_file_dir
1: CREATE OR REPLACE DIRECTORY
ฺ m n t AS '/home/oracle/'
Directory created. d gp Stud

b i a@ this
i a (ja use
Grant succeeded.
A b to
o
ni succeeded.
ntoGrant
e A
J o s Grant succeeded.
Connected.
Session altered.
Session altered.
Table created.
Table created.
...
Table created.

(continued)
Creating constraints ...
Table altered.
...
Table altered.

n s e
i ce
el
specify path for data files as parameter 2:
a b l
specify path for log files as parameter 3:
fer
a n s
-tr
o n
an
Looking for indexes that could slow down load ...
s
no rows selected
) h a deฺ
r ฺ e s Gui
loading TIMES using: p ฺ mi dent
d g Stu
/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/time_v
3.ctl
b i a@ this
3.dat
i a (ja use
A b to
/home/oracle/time_v3.log
ni o
ntoSQL*Loader: Release 11.1.0.6.0 - Production on Thu Mar 20 15:22:23
e A 2008
J o s Copyright (c) 1982, 2007, Oracle. All rights reserved.
Save data point reached - logical record count 1000.
Load completed - logical record count 1826.
loading COUNTRIES using:

/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/coun_v
3.ctl
3.dat
/home/oracle/coun_v3.log
SQL*Loader: Release 11.1.0.6.0 - Production on Thu Mar 20 15:22:24

2008

(continued)
loading CUSTOMERS using:

/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/cust_v
3.ctl
/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/cust1v
3.dat
/home/oracle/cust1v3.log

2008

n s e
i ce
el
a b l
fer
a n s
-tr
o n
s an
) h a deฺ
loading PRODUCTS using:
r ฺ e s Gui
3.ctl p ฺ mi dent
/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/prod_v
d g Stu
/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/prod1v
3.dat
b i a@ this
i a (ja use
/home/oracle/prod1v3.log
A
SQL*Loader: to 11.1.0.6.0 - Production on Thu Mar 20 15:22:24
b Release
i o
ton
2008
A n Copyright (c) 1982, 2007, Oracle. All rights reserved.

o s e
J
loading PROMOTIONS using:

/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/prom_v
3.ctl
/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/prom1v
3.dat
/home/oracle/prom1v3.log

2008

...

(continued)
loading CHANNELS using:

/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/chan_v
3.ctl
3.dat
/home/oracle/chan_v3.log

2008

n s e
i ce
b l el
fer a
a n s
loading SALES using:
o n -tr
3.ctl
s an
/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/sale_v
h a deฺ
/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/sale1v
)
3.dat
r ฺ e s Gui
/home/oracle/sale1v3.log
p ฺ mi dent
d g S-tuProduction on Thu Mar 20 15:22:25
SQL*Loader: Release 11.1.0.6.0
2008
b i a@ this
Copyright a
i (ja1982,u2007,
(c) se Oracle. All rights reserved.
A b o
treached
n i
Save o data point - logical record count 100000.
nt o...
e A Save data point reached - logical record count 900000.
J o s Load completed - logical record count 916039.
loading COSTS using external table
Table created.
82112 rows created.
loading additonal SALES using:

/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/dmsal_
v3.ctl
v3.dat
/home/oracle/dmsal_v3.log

(continued)
2008


...
loading SUPPLEMENTARY DEMOGRAPHICS using:

/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/dem_v3
.ctl
n s e
i ce
el
/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/dem1v3
.dat
a b l
er
/home/oracle/dem1v3.log
s f
n 15:22:38
SQL*Loader: Release 11.1.0.6.0 - Production on Thu Mar a
- t r 20
2008
n on
s a reserved.
) h a
i d eฺ
... i r ฺes nt Gu
Save data point reached - p

g ฺm urecord
logical de count 4500.
d t
Scount 4500.
i a @ record
i s
ab se th
Load completed - logical
( j
b ia to u
oA
Commit complete.
o n i
nt
eA Enabling constraints ...
Jos Table altered.
...
Table altered.
Creating additional indexes ...
Index created.
...
Index created.

(continued)
Create dimensions ...
Dimension created.
Commit complete.
no rows selected
Dimension created.
n s e
i ce
b l el
fer a
a n s
no rows selected n -tr
o
s an
Dimension created.
) h a deฺ
r ฺ e s Gui
ฺ i ent
mcompleted.
p
g Stud
d
b i a@ this
i a (ja use
no rows selected
A b to
i o
ton
Dimension created.
A n
o s e PL/SQL procedure successfully completed.
J
no rows selected
Dimension created.
no rows selected
Creating MVs as tables ...
View created.
Table created.

(continued)
Table created.
Index created.
Index created.
Index created.
Index created.
n s e
i ce
Creating materialized views ...
b l el
fer a
a n s
Materialized view created.
n -tr
o
s an
) h a deฺ
r ฺ e s Gui
Creating comments ...
p ฺ mi dent
Comment created. d g Stu
b i a@ this
...
i a (ja use
A b to
ni o
ntoComment created.
e A
J o s Creating OLAP metadata ...
-
<<<<< CREATE CATALOG sh_cat for Sales History >>>>>
Catalog Dropped
CWM Collect Garbage
-
<<<<< CREATE the Sales CUBE >>>>>
Sales amount, Sales quantity
<TIMES CHANNELS PRODUCTS CUSTOMERS PROMOTIONS >
Drop SALES_CUBE prior to recreation
Cube Dropped
Add dimensions -
to SALES_CUBE and map the foreign keys
Create measures -
for SALES_CUBE and map to columns in the fact table
Set default aggregation method -
to SUM for all measures over TIME
Add SALES_CUBE to the catalog
SALES_CUBE successfully added to sh_cat

(continued)
-
<<<<< CREATE the Cost CUBE >>>>>
Unit Cost, Unit Price < TIMES PRODUCTS CHANNELS PROMOTIONS >
Drop COST_CUBE prior to recreation
Cube Dropped
Add dimensions -
to COST_CUBE and map the foreign keys
Create measures -
for COST_CUBE and map to columns in the fact table
Add COST_CUBE to the catalog
COST_CUBE successfully added to sh_cat
-
<<<<< TIME DIMENSION >>>>>
n s e
i ce
el
a b l
fer
a n s
Level Attributes - name, display name, description n - tr
o
s an
) h a deฺ
r ฺ e s Gui
p ฺ mi dent
- Period Number of Days created
- Period End Date created
d g Stu
i a@ this
- Time dimension
b
i a (ja use
- Long description
- Day name
Ab to
- Calendar month description
o n io - Calendar quarter description

nt - Fiscal month description
e A - Fiscal quarter description
J o s - Short Description
- Day name
- Calendar quarter description
- Fiscal month description
- Fiscal quarter description
- Time Span
- Days in calendar month
- Days in calendar quarter
- Days in calendar year
- Days in fiscal month
- Days in fiscal quarter
- Days in fiscal year
- End Date
- End of calendar month
- End of calendar quarter
- End of calendar year
- End of fiscal month
- End of fiscal quarter
- End of fiscal year
-

(continued)
<<<<< CUSTOMERS DIMENSION >>>>>

n s e
i ce
el
a b l
fer
a n s
n - tr
o
s an
h
- Other Customer Information created
) a deฺ
r ฺ e s Gui
- Long Description
- Short Descriptionp ฺ mi dent
- d g Stu
i a@ this
<<<<< PRODUCTS DIMENSION >>>>>
b
i a (ja use
Ab to
o n io - default calculation hierarchy

nt - default display hierarchy
e A Level Attributes - name, display name, description
J o s Drop dimension attributes prior to re-creation

- Long Description
- Short Description
-
<<<<< PROMOTIONS DIMENSION >>>>>

(continued)
- Long Description
- Short Description
-

n s e
i ce
el
a b l
- Long Description
fer
- Short Description
a n s
-
<<<<< FINAL PROCESSING >>>>> n -tr
o
s an
) h a deฺ
r ฺ e s Gui
Commit complete. p ฺ mi dent
d g Stu
b i a@... this
i a (ja use
gathering statistics
A b
PL/SQL procedure tosuccessfully completed.
i o
nton
e A PL/SQL procedure successfully completed.
J o s SQL> SQL> Disconnected from Oracle Database 11g Enterprise Edition

[oracle@edrsr33p1-orcl SQL_Access_Advisor]$
--------------------------------------------------------------
#!/bin/bash
export

(continued)
#
# Cleanup ADDM snapshot settings
#
sqlplus -s /NOLOG <<EOF >> /tmp/cleanup_dina.log 2>&1
connect / as sysdba
rem -- change INTERVAL setting to 30 minute

execute dbms_workload_repository.modify_snapshot_settings(interval
=> 60);
rem -- change ADDM sensitiveness back to normal

exec
n s e
i ce
el
dbms_advisor.set_default_task_parameter('ADDM','DB_ACTIVITY_MIN',300
);
a b l
fer
connect sh/sh
a n s
drop index sales_time_idx; n - tr
o
create bitmap index sales_time_bix
s an
on sales(time_id)
) h a deฺ
tablespace example
r ฺ e s Gui
p ฺ mi dent
local nologging compute statistics;
d g Stu
EOF
b i a@ this
#
i a (ja use
# Cleanup
# io
to
Ab sql profile
n
tosqlplus -s /NOLOG <<EOF > /tmp/cleanup_dina.log 2>&1
A n
o s e connect / as sysdba
J
set head off
set timing off
set feedback off;
set pagesize 0
spool /tmp/drop_dyn.sql;
select q'#connect / as sysdba;#' from dual;
select q'#execute dbms_sqltune.drop_sql_profile('#' || name || q'#')

;#'
from dba_sql_profiles ;
select q'#execute dbms_advisor.delete_task('#' || task_name || q'#')

;#'
from user_advisor_tasks
where CREATED > SYSDATE-(1/24);
select q'#connect system/oracle;#' from dual;

(continued)
select q'#execute dbms_advisor.delete_task('#' || task_name || q'#')
;#'
from user_advisor_tasks
where CREATED > SYSDATE-(1/24);
spool off
@/tmp/drop_dyn.sql
EOF
cp /home/oracle/solutions/SQL_Access_Advisor/sh/*
$ORACLE_HOME/demo/schema/sales_history
n s e
i ce
b l el
fer a
sqlplus -s /NOLOG <<EOF
a n s
set echo on
o n -tr
connect / as sysdba
s an
) h a deฺ
revoke dba from sh;
r ฺ e s Gui
p ฺ
@sh_main sh example temp oraclemi dent
d g Stu
/home/oracle/ v3 a@ is
b i t h
EOF
i a (ja use
A b to
i o
nton
e A
J o s

Practice 11-2: Using SQL Access Advisor
The following scenario illustrates the types of recommendations that can be made by
SQL Access Advisor. The scenario also uses the SQL Performance Analyzer to prove
that recommendations made by SQL Access Advisor are good.
1) From a terminal session connected as the oracle user, execute the
sqlaccessadv_setup.sh script. This script generates the necessary data that

you use throughout this lab. In particular, it generates the SQL Tuning Set that is used
to represent the workload you want to analyze.
./sqlaccessadv_setup.sh
[oracle@edrsr33p1-orcl SQL_Access_Advisor]$ ./sqlaccessadv_setup.sh
SQL*Plus: Release 11.1.0.6.0 - Production on Tue Mar 18 21:14:49 n s e

ice
el
2008
a b l
fe r
a n s
Connected to:
o n -tr
an
s
Production
) h a deฺ
s Gui
ฺ
r
i ente
p ฺ mSQL>
SQL> SQL> d g StudSQL> SQL> SQL> SQL> Connected.
SQL> SQL> SQL> SQL> SQL> SQL>
Grant succeeded. ia@ is

b
ja use t h
SQL> SQL> a (
b i to
A
User altered.
o
ni SQL> Connected.
toSQL>
A n SQL> SQL> SQL> SQL> SQL> SQL> SQL> SQL> SQL> SQL> 2 3 4 5
o s e 6 7 8 9 10 11 12 13 14 15 16 17 18 19
J 20 21 22 23 24
SQL> SQL> SQL> SQL> SQL> SQL> SQL> DROP TABLE temp_table purge
*
ERROR at line 1:
SQL>
Table created.
SQL> SQL> SQL> SQL> SQL> SQL> SQL>

System altered.
SQL> BEGIN dbms_sqltune.drop_sqlset('SQLSET_MY_SQLACCESS_WORKLOAD');

END;
*
ERROR at line 1:

Practice 11-2: Using SQL Access Advisor (continued)
ORA-13754: "SQL Tuning Set" "SQLSET_MY_SQLACCESS_WORKLOAD" does not
exist for user "SH".
ORA-06512: at "SYS.DBMS_SQLTUNE_INTERNAL", line 8406
ORA-06512: at "SYS.DBMS_SQLTUNE", line 2949
ORA-06512: at line 1
SQL> SQL> drop table tempjfv purge

*
ERROR at line 1:
SQL> SQL>
Table created.
n s e
SQL> SQL> SQL> 2 3 4 5 6 7
i ce
b l el
SQL> SQL> drop table customersjfv purge
fer a
*
a n s
ERROR at line 1:
o n -tr
an
h s
a deฺ
SQL>
e )
s Gui
Table created. r ฺ
i ent
p
3 g 4
ฺ m 5ud 6
SQL> SQL> SQL> 2
16 @ d S t 20 217 228 239 2410 11 12
13
27
14
28
15
29 b
30i
17 18
a 31 t32
h is 33 34 35 36 37 38
19 25
39
26
40
41 42 43(ja44 se 46 47
45 48 49
b i a
PL/SQL procedure
to u
successfully completed.
o A
t o niCOUNT(*)
SQL> SQL> SQL>
An
e
Jos
----------
5
1 row selected.
SQL> SQL> SQL> SQL> SQL> SQL> SQL> SQL>

SQL>
SQL> SQL>
SQL> SQL>
SQL> SQL>
LAST_ANAL
---------
18-MAR-08

18-MAR-08
18-MAR-08
18-MAR-08
22-AUG-07
18-MAR-08
18-MAR-08
18-MAR-08
18-MAR-08
18-MAR-08
18-MAR-08
18-MAR-08
18-MAR-08
14 rows selected.
n s e
SQL> SQL> SQL> SQL> Disconnected from Oracle Database 11g Enterprise
i ce
b l el
fer a
a n s
o n -tr
an
--------------------------------------------------------------
#!/bin/bash
h s
a deฺ
e )
s Gui
r ฺ
i ent
cd /home/oracle/solutions/SQL_Access_Advisor
p ฺ m
export ORACLE_SID=orcl g
d S tud
b i a@ this
export bia
(ja use
A to
ni o
ntol/bin:/usr/X11R6/bin:/usr/java/jdk1.5.0_11/bin:/bin
e A
J o s sqlplus / as sysdba <<FIN!
SET ECHO ON
SET FEEDBACK 1
SET NUMWIDTH 10
SET LINESIZE 8000
SET TRIMSPOOL ON
SET TAB OFF
SET PAGESIZE 100
SET LONG 1000
CONNECT / AS SYSDBA
grant dba to sh;
connect sh/sh
set serveroutput on size 32768;

set echo on;

variable norecs number;
Rem Clean up
declare
name varchar2(30);
cursor name_cur1 is
select task_name from user_advisor_templates
where task_name like '%SQLACCESS%';
begin
----------------------------------------------------------------
-----------
-- Get rid of templates, tasks and workloads.
----------------------------------------------------------------
n s e
-----------
i ce
b l el
open name_cur1;
fer a
loop
a n s
fetch name_cur1 into name;
o n -tr
an
exit when name_cur1%NOTFOUND;
h s
a deฺ
)
s Gui
dbms_advisor.update_task_attributes(name,null,null,'FALSE','FALSE');
e
r ฺ
mi dent
dbms_advisor.delete_task(name);
end loop;
p
g Stuฺ
close name_cur1;@d
b i a t h is
end;
i a (ja use
/
A b to
i o
nton
e A
J o s Rem make a temp table
DROP TABLE temp_table purge;

CREATE TABLE temp_table AS SELECT * FROM SYS.WRI\$_ADV_SQLW_STMTS
WHERE NULL IS NOT NULL;
Rem create a large number of pseudo-random (repeatable) queries in

the temporary table

execute dbms_sqltune.drop_sqlset('SQLSET_MY_SQLACCESS_WORKLOAD');
drop table tempjfv purge;
create table tempjfv (c number, d varchar2(1000));

begin
insert into tempjfv values(-
i,'aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa');
end loop;
commit;
end;
/
n s e
drop table customersjfv purge;
i ce
create table customersjfv as select * from customers;
b l el
fe r a
DECLARE
a n s
sql_stmt varchar2(2000);
o n -tr
an
sqlsetname VARCHAR2(30);
sqlsetcur dbms_sqltune.sqlset_cursor;
refid NUMBER;
h s
a deฺ
k NUMBER := 0;
e )
s Gui
r ฺ
mi dent
num_queries NUMBER := 500;
BEGIN
p ฺ
g S2tu*/ ch.channel_class, c.cust_city,
sql_stmt := 'SELECT @ /* d
b i
t.calendar_quarter_desc,a t h is
QueryJFV
SUM(s.amount_sold) sales_amount FROM
sh.sales s, ( j a
sh.times e
sAND s.cust_id = c.cust_id AND ch
t, sh.customers c, sh.channels WHERE
s.time_idb i a
= t.time_id u
toc.cust_state_province = ''CA'' AND ch.channel_desc
s.channel_id =
A
io(''Internet'',''Catalog'') AND t.calendar_quarter_desc IN
ch.channel_id AND
o n
in
t (''1999-01'',''1999-02'') GROUP BY ch.channel_class, c.cust_city,
A n
o s e t.calendar_quarter_desc';
J insert into temp_table values(1,1,NULL,0,'SH','Access
Advisor','Workload',0,0,0,0,1,100,2,to_date('02-FEB-
2007'),3,0,sql_stmt,1);
sql_stmt := 'SELECT /* QueryJFV 3 */ ch.channel_class, c.cust_city,

t.calendar_quarter_desc, SUM(s.amount_sold) sales_amount FROM
sh.sales s, sh.times t, sh.customers c, sh.channels ch WHERE
s.time_id = t.time_id AND s.cust_id = c.cust_id AND s.channel_id =
ch.channel_id AND c.cust_state_province = ''CA'' AND ch.channel_desc
in (''Internet'',''Catalog'') AND t.calendar_quarter_desc IN
(''1999-03'',''1999-04'') GROUP BY ch.channel_class, c.cust_city,
t.calendar_quarter_desc';
insert into temp_table values(1,1,NULL,0,'SH','Access

2007'),3,0,sql_stmt,1);

sql_stmt := 'SELECT /* QueryJFV 4 */ c.country_id, c.cust_city,
c.cust_last_name FROM sh.customers c WHERE c.country_id in (52790,
52798) ORDER BY c.country_id, c.cust_city, c.cust_last_name';

2007'),3,0,sql_stmt,1);
sql_stmt := 'select /* func_indx */ count(*) from tempjfv where

abs(c)=5';

2007'),3,0,sql_stmt,1);
n s e
sql_stmt := 'SELECT /* QueryJFV 5 */ * FROM sh.customersjfv WHERE
i ce
cust_state_province = ''CA''';
b l el
fer a
a n s
2007'),3,0,sql_stmt,1);
o n -tr
s an
h
sqlsetname := 'SQLSET_MY_SQLACCESS_WORKLOAD';
) a deฺ
r ฺ e G ui
s 'Generated
m i
dbms_sqltune.create_sqlset(sqlsetname,
n t STS');
OPEN sqlsetcur FOR

d g pฺ tude
SELECT
i a @ his Ssql_text, null, null, username,
module, ( j ab se t
SQLSET_ROW(null,null,
b ia to u action, elapsed_time, cpu_time, buffer_gets,

i oA
disk_reads,
n
ntooptimizer_cost, null,0,rows_processed, 0, executions, 0,
e A
J o s priority, command_type,
to_char(last_execution_date,'yyyy-mm-
dd/hh24:mi:ss'),
0,0,NULL,0,NULL,NULL
)
FROM temp_table;
dbms_sqltune.load_sqlset(sqlsetname, sqlsetcur);
END;
/
SELECT COUNT(*) FROM

TABLE(DBMS_SQLTUNE.SELECT_SQLSET('SQLSET_MY_SQLACCESS_WORKLOAD'));
Rem Cleanup anything left behind

execute dbms_advisor.delete_task('%');
execute dbms_advisor.delete_sqlwkld('%');
EXECUTE DBMS_STATS.UNLOCK_SCHEMA_STATS('SH');
execute dbms_stats.gather_schema_stats(ownname => 'SH',

estimate_percent=> DBMS_STATS.AUTO_SAMPLE_SIZE, method_opt => 'FOR

ALL COLUMNS SIZE AUTO', degree => 4 );
select distinct last_analyzed from dba_tab_statistics where

owner='SH';
REM EXECUTE DBMS_STATS.LOCK_SCHEMA_STATS('SH');
exit;
n s e
FIN!
ice
b l el
fe
2) Using Enterprise Manager, create a SQL Access Advisor tuning task based on the ra
captured workload held in the SH.SQLSET_MY_ACCESS_WORKLOAD SQL a n s
tuning set using the SQLACCESS_WAREHOUSE template.
o n -tr
n
a) Connect to Enterprise Manager Database Control asathe sh user (password: sh).
On the Home page, click the Advisor Centralhlink
s
a in the ฺ Links section.
eRelated
)
s Advisors i d
u link. Then on the SQL
b) On the Advisor Central page, clickithe r ฺeSQL t G
Advisors page, click the SQLpAccess
g ฺm uAdvisor
d en link.
d S t Options from a previously saved Task
c) On the Initial Options @ s
page, select Inherit
a selectththei SQLACCESS_WAREHOUSE template. After
ithen
or Template,ja b
and
i a
this is done, se
( click Continue.
u
OnA
d) io
b to
the Workload Source page, select Use an existing SQL Tuning Set and enter
n
Anto SH.SQLSET_MY_SQLACCESS_WORKLOAD in the SQL Tuning Set field.
(This SQL Tuning Set was generated earlier. It represents a warehouse workload
e
Jos that you want to analyze.) Click Next.
e) On the Recommendation Options page, ensure that all possible access structures
are selected, and that Comprehensive is selected. After this is done, click Next.
f) On the Schedule page, enter MY_SQLACCESS_TASK in the Task Name field.
Select the first Time Zone from the provided list (Click the torch icon.). After this
is done, click Next.
g) On the Review page, click Submit.
h) Back to the Advisor Central page, click Refresh until the Status of your task is
COMPLETED.
i) After this is done, click the MY_SQLACCESS_TASK link in the Results table.
3) After this is done, investigate the proposed recommendations:
a) Back to the Advisor Central page, click the MY_SQLACCESS_TASK link in the
Results table. The task should have COMPLETED as the status.

b) This takes you to the Results page. From this page, you can see the potential
benefit of implementing the SQL Access Advisor recommendations on the
workload. There should be a huge difference between the original and the new
costs. Click the Recommendation subtab.
c) On the Recommendations subtab, you can see the high-level overview of the
recommendations. Basically, all possible types of recommendations were

generated for this workload (Indexes, Materialized Views, Materialized View
Logs, Partitions, and Others).
d) Ensure that all recommendations are selected, and click the Recommendation
Details button. This takes you to the Details page, where you can see more details
about each of the recommendations, as well as the corresponding SQL statements
from the workload that are affected by these recommendations. You should see e
the following recommendations: n s
ice
- Partition CUSTOMERS table.
b l el
- Create four materialized view log.
fe ra
- Create a materialized view. s
a n
- Create one bitmap index.
- Create a function-based index. o n -tr
- Create a B*-tree index on multiple columns.
s an
) h a deฺ
e) Click OK.
r ฺ e s Gui
p ฺ mi dent What happens, and why?
4) Try to implement the generated recommendations.
d
a) Back to the Recommendationsg subtab,
S tuclick the Schedule Implementation button.
i a@ this page, a warning is displayed indicating that the
b) On the SchedulebImplementation
wizard will
a ja try touimplement
(not se its recommendations because some of them are
b i t o
i o Aimportant changes
very that should be looked at closely by the administrator.
n
Antoc) Click the Show SQL button to look at the script you could use to implement all
recommendations. In fact, you already created this script and you will use it later
e
Jos in this lab. After you review the script, click Done.
d) Back to the Schedule Implementation page, click Cancel.
e) Click the Advisor Central locator link at the top of the “Results for Task” page.
f) On the Advisor Central page, select the SQL Access Advisor
MY_SQLACCESS_TASK task and click Delete.
g) On the Information page, click Yes.
5) Use Enterprise Manager to verify the performance improvement if you implement the
recommendations mentioned.
a) Click the Database tab at the top-right corner and then the “Software and
Support” tab. On the “Software and Support” tabbed page, click the SQL
Performance Analyzer link. You want to prove that implementing the
recommendations is beneficial.
b) On the SQL Performance Analyzer page, click the Guided Workflow link.

c) On the Guided Workflow page, click the Execute icon on the line corresponding
to step 1.
d) On the Create SQL Performance Analyzer Task page, enter MY_SPA_TASK in
the SQL Performance Analyzer Task Name field. Then, enter
SH.SQLSET_MY_SQLACCESS_WORKLOAD in the SQL Tuning Set Name field.
After this is done, click Create.

e) Back to the Guided Workflow page, click the Execute icon for step 2.
f) On the Create Replay Trial page, enter MY_SQL_REPLAY_BEFORE in the
Replay Trial Name field, and ensure that you select the Trial environment
established check box. Then, click Submit.
g) Wait on the Guided Workflow page until step 2 is completed.
n s e
h) From your terminal session, connect as the sh user (password: sh) in the i ce
SQL*Plus session, and execute the implement.sql script. This script is a b l el
fe r a
precreated script corresponding to the recommendations previously generated by
your SQL Access Advisor session. a n s
o n -tr
an
cd /home/oracle/solutions/SQL_Access_Advisor
[oracle@edrsr33p1-orcl SQL_Access_Advisor]$ ls
implement.sql revert.sh sqlaccessadv_setup.sh
h s
a deฺ
)
s Gui
[oracle@edrsr33p1-orcl SQL_Access_Advisor]$ sqlplus sh/sh
e
r ฺ
i ent on Tue Mar 18 21:57:25
SQL*Plus: Release 11.1.0.6.0ฺm - Production
2008
d p
g Stud
Copyright (c) 1982,b i a@
2007,th is
( j a s e Oracle. All rights reserved.
b ia to u
i o ADatabase
Connected
n
to:
ntoOracle 11g Enterprise Edition Release 11.1.0.6.0 -
e A Production
J o s With the Partitioning, Oracle Label Security, OLAP, Data Mining

SQL> @implement
SQL>
SQL> Rem
SQL> Rem Creating new partitioned table
SQL> Rem
SQL> CREATE TABLE "SH"."CUSTOMERS1"
2 ( "CUST_ID" NUMBER,
3 "CUST_FIRST_NAME" VARCHAR2(20),
4 "CUST_LAST_NAME" VARCHAR2(40),
5 "CUST_GENDER" CHAR(1),
6 "CUST_YEAR_OF_BIRTH" NUMBER(4,0),
7 "CUST_MARITAL_STATUS" VARCHAR2(20),
8 "CUST_STREET_ADDRESS" VARCHAR2(40),
9 "CUST_POSTAL_CODE" VARCHAR2(10),
10 "CUST_CITY" VARCHAR2(30),
11 "CUST_CITY_ID" NUMBER,
12 "CUST_STATE_PROVINCE" VARCHAR2(40),
13 "CUST_STATE_PROVINCE_ID" NUMBER,

14 "COUNTRY_ID" NUMBER,
15 "CUST_MAIN_PHONE_NUMBER" VARCHAR2(25),
16 "CUST_INCOME_LEVEL" VARCHAR2(30),
17 "CUST_CREDIT_LIMIT" NUMBER,
18 "CUST_EMAIL" VARCHAR2(30),
19 "CUST_TOTAL" VARCHAR2(14),
20 "CUST_TOTAL_ID" NUMBER,
21 "CUST_SRC_ID" NUMBER,
22 "CUST_EFF_FROM" DATE,
23 "CUST_EFF_TO" DATE,
24 "CUST_VALID" VARCHAR2(1)
25 ) PCTFREE 10 PCTUSED 40 INITRANS 1 MAXTRANS 255 NOCOMPRESS
NOLOGGING
26 TABLESPACE "EXAMPLE"
27 PARTITION BY RANGE ("CUST_ID") INTERVAL( 3000) ( PARTITION
VALUES LESS THAN (3000)
n s e
28 );
i ce
b l el
Table created.
fer a
SQL>
a n s
SQL> Rem
o n -tr
an
SQL> Rem Copying comments to new partitioned table
SQL> Rem
SQL>
h s
a deฺ
COMMENT ON COLUMN "SH"."CUSTOMERS1"."CUST_ID" IS 'primary key';
e )
s Gui
Comment created. r ฺ
i ent
p ฺ m d
SQL>
d g S t u
b i a@ "SH"."CUSTOMERS1"."CUST_FIRST_NAME"
SQL> COMMENT ON COLUMN
name of the customer'; t h is IS 'first
ja e
ia ( to us
Comment bcreated.
n i oA
ntoSQL> COMMENT ON COLUMN "SH"."CUSTOMERS1"."CUST_LAST_NAME" IS 'last
SQL>
e A
J o s name of the customer';
Comment created.
SQL>
SQL> COMMENT ON COLUMN "SH"."CUSTOMERS1"."CUST_GENDER" IS 'gender;
low cardinality attribute';
Comment created.
SQL>
SQL> COMMENT ON COLUMN "SH"."CUSTOMERS1"."CUST_YEAR_OF_BIRTH" IS
'customer year of birth';
Comment created.
SQL>
SQL> COMMENT ON COLUMN "SH"."CUSTOMERS1"."CUST_MARITAL_STATUS" IS
'customer marital status; low cardinality attribute';
Comment created.

SQL>
SQL> COMMENT ON COLUMN "SH"."CUSTOMERS1"."CUST_STREET_ADDRESS" IS
'customer street address';
Comment created.
SQL>
SQL> COMMENT ON COLUMN "SH"."CUSTOMERS1"."CUST_POSTAL_CODE" IS
'postal code of the customer';
Comment created.
SQL>
SQL>
SQL> COMMENT ON COLUMN "SH"."CUSTOMERS1"."CUST_CITY" IS 'city where
n s e
the customer lives';
i ce
b l el
Comment created.
fer a
SQL>
a n s
o n -tr
SQL> COMMENT ON COLUMN "SH"."CUSTOMERS1"."CUST_STATE_PROVINCE" IS
an
'customer geography: state or province';
Comment created.
h s
a deฺ
e )
s Gui
SQL> r ฺ
i ent
p ฺ m
SQL> COMMENT ON COLUMN "SH"."CUSTOMERS1"."COUNTRY_ID"
d IS 'foreign
d
key to the countries tableg t u
(snowflake)';
S
@ s
Comment created.bia
j a e thi
SQL> b ia ( to us
i o A mainONphone
SQL> COMMENT
n
COLUMN "SH"."CUSTOMERS1"."CUST_MAIN_PHONE_NUMBER" IS
nto'customer number';
e A
J o s Comment created.
SQL>
SQL> COMMENT ON COLUMN "SH"."CUSTOMERS1"."CUST_INCOME_LEVEL" IS
'customer income level';
Comment created.
SQL>
SQL> COMMENT ON COLUMN "SH"."CUSTOMERS1"."CUST_CREDIT_LIMIT" IS
'customer credit limit';
Comment created.
SQL>
SQL> COMMENT ON COLUMN "SH"."CUSTOMERS1"."CUST_EMAIL" IS 'customer
email id';
Comment created.
SQL>

SQL> COMMENT ON TABLE "SH"."CUSTOMERS1" IS 'dimension table';
Comment created.
SQL>
SQL> Rem
SQL> Rem Copying constraints to new partitioned table

SQL> Rem
SQL> ALTER TABLE "SH"."CUSTOMERS1" ADD CONSTRAINT "CUSTOMERS_PK1"
PRIMARY KEY ("CUST_ID")
2 USING INDEX PCTFREE 10 INITRANS 2 MAXTRANS 255 NOLOGGING
COMPUTE STATISTICS
3 TABLESPACE "EXAMPLE" ENABLE NOVALIDATE;
Table altered.
n s e
SQL>
i ce
SQL> Rem
b l el
SQL> Rem Copying referential constraints to new partitioned table
SQL> Rem
fer a
SQL> ALTER TABLE "SH"."CUSTOMERS1" ADD CONSTRAINT
a n s
"CUSTOMERS_COUNTRY_FK1" FOREIGN KEY ("COUNTRY_ID")
o n -tr
an
2 REFERENCES "SH"."COUNTRIES" ("COUNTRY_ID") ENABLE
NOVALIDATE;
h s
a deฺ
Table altered.
e )
s Gui
r ฺ
i ent
SQL>
p ฺ m d
SQL> Rem
d g S t u
SQL> Rem b i a@ tto
SQL> Rem Copying indexes
h isnew partitioned table
SQL> CREATE (
a ja uINDEX
BITMAP se "SH"."CUSTOMERS_GENDER_BIX1" ON
i o
b 10 tINITRANS
"SH"."CUSTOMERS1" ("CUST_GENDER")
i o A
2 PCTFREE 2 MAXTRANS 255 NOLOGGING COMPUTE STATISTICS
to n 3 TABLESPACE "EXAMPLE" LOCAL;
A n
o s e Index created.
J SQL>
SQL> Rem
SQL> Rem Copying object grants to new partitioned table
SQL> Rem
SQL> GRANT SELECT ON "SH"."CUSTOMERS1" TO "BI";
Grant succeeded.
SQL>
SQL> Rem
SQL> Rem Populating new partitioned table with data from original
table
SQL> Rem
SQL> INSERT /*+ APPEND */ INTO "SH"."CUSTOMERS1"
2 SELECT * FROM "SH"."CUSTOMERS";
55500 rows created.
SQL> COMMIT;

Commit complete.
SQL>
SQL> begin
2 dbms_stats.gather_table_stats('"SH"', '"CUSTOMERS1"', NULL,
dbms_stats.auto_sample_size);
3 end;
4 /
SQL>
SQL> Rem
SQL> Rem Renaming tables to give new partitioned table the original
table name
n s e
SQL> Rem
i ce
SQL> ALTER TABLE "SH"."CUSTOMERS" RENAME TO "CUSTOMERS11";
b l el
Table altered.
fer a
a n s
SQL> ALTER TABLE "SH"."CUSTOMERS1" RENAME TO "CUSTOMERS";
o n -tr
Table altered.
s an
) h a deฺ
SQL>
r ฺ e s Gui
SQL> Rem
ฺ
SQL> Rem Revalidating dimensionsm i for eusent with new partitioned table
SQL> Rem p
dg Stu d
SQL> ALTER DIMENSION@
b i a t h is
"SH"."CUSTOMERS_DIM" COMPILE;
i a (ja use
Dimension altered.
SQL> A
b to
n i o
ntoSQL>
e A SQL> CREATE MATERIALIZED VIEW LOG ON
J o s 2 "SH"."CUSTOMERS"
3 WITH ROWID,
SEQUENCE("CUST_ID","CUST_CITY","CUST_STATE_PROVINCE")
4 INCLUDING NEW VALUES;
Materialized view log created.
SQL>
SQL> CREATE MATERIALIZED VIEW LOG ON
2 "SH"."CHANNELS"
3 WITH ROWID,
SEQUENCE("CHANNEL_ID","CHANNEL_DESC","CHANNEL_CLASS")
SQL>
2 "SH"."TIMES"
3 WITH ROWID, SEQUENCE("TIME_ID","CALENDAR_QUARTER_DESC")

SQL>
2 "SH"."SALES"
3 WITH ROWID,
SEQUENCE("CUST_ID","TIME_ID","CHANNEL_ID","AMOUNT_SOLD")
SQL>
SQL> CREATE MATERIALIZED VIEW "SH"."MV_01DF0000"
2 REFRESH FAST WITH ROWID
3 ENABLE QUERY REWRITE
n s e
4 AS SELECT SH.CUSTOMERS.CUST_STATE_PROVINCE C1,
i ce
SH.CUSTOMERS.CUST_CITY C2, SH.CHANNELS.CHANNEL_CLASS
b l el
5 C3, SH.CHANNELS.CHANNEL_DESC C4, SH.TIMES.CALENDAR_QUARTER_DESC
C5, SUM("SH"."SALES"."AMOUNT_SOLD")
fer a
n
6 M1, COUNT("SH"."SALES"."AMOUNT_SOLD") M2, COUNT(*) M3 FROM
a s
SH.CUSTOMERS,
n -tr
o
an
7 SH.CHANNELS, SH.TIMES, SH.SALES WHERE SH.SALES.CHANNEL_ID =
SH.CHANNELS.CHANNEL_ID
h s
a deฺ
8 AND SH.SALES.TIME_ID = SH.TIMES.TIME_ID AND SH.SALES.CUST_ID =
SH.CUSTOMERS.CUST_ID
e )
s Gui
r ฺ
mi dent
9 AND (SH.TIMES.CALENDAR_QUARTER_DESC IN ('1999-04', '1999-03',
'1999-02'
p ฺ
g Stu
d
10 , '1999-01')) AND (SH.CHANNELS.CHANNEL_DESC IN ('Internet',
'Catalog'
b i a@ this
11 )) AND (SH.CUSTOMERS.CUST_STATE_PROVINCE = 'CA') GROUP BY
i a (ja use
SH.CUSTOMERS.CUST_STATE_PROVINCE,
b to
12 SH.CUSTOMERS.CUST_CITY, SH.CHANNELS.CHANNEL_CLASS,
A
o n io
SH.CHANNELS.CHANNEL_DESC,
13 SH.TIMES.CALENDAR_QUARTER_DESC;
nt
e A
J o s Materialized view created.
SQL>
SQL> begin
2
dbms_stats.gather_table_stats('"SH"','"MV_01DF0000"',NULL,dbms_stats
.auto_sample_size);
3 end;
4 /
SQL>
SQL> CREATE BITMAP INDEX "SH"."CUSTOMERSJFV_IDX_01DF0000"
2 ON "SH"."CUSTOMERSJFV"
3 ("CUST_STATE_PROVINCE")
4 COMPUTE STATISTICS;
Index created.
SQL>

SQL> CREATE INDEX "SH"."TEMPJFV_IDX_01DF0001"
2 ON "SH"."TEMPJFV"
3 (ABS("C"))
Index created.
SQL>
SQL> CREATE INDEX "SH"."CUSTOMERS_IDX_01DF0002"
2 ON "SH"."CUSTOMERS"
3 ("COUNTRY_ID","CUST_CITY","CUST_LAST_NAME")
Index created.
SQL>
n s e
SQL> exit
ice
b l el
fe r a
a n s
o n -tr
i) Back to your Guided Workflow page, click the Execute a nicon corresponding to
step 3.
h s
a deฺ
e )
s Gui
j) On the Create Replay Trial page, enter
i r ฺ MY_SQL_REPLAY_AFTER
n t in the Replay
Trial Name field. Ensure that you
p m
ฺ ude
select the Trial environment established check
box, and click Submit. d g S t
@ s
k) Wait until step b
j a 3 is
e thi
iacompleted.
l) Backb toia
(
your Guided u s
Workflow Enterprise Manager page, click the Execute icon
A
corresponding
t o
to step 4.
n i o
Antom) On the Run Replay Trial Comparison page, ensure that you create a comparison
e between MY_SQL_REPLAY_BEFORE and MY_SQL_REPLAY_AFTER. Click
Jos Submit.
n) Wait until step 4 is completed.
o) Back to your Guided Workflow Enterprise Manager page, click the Execute icon
corresponding to step 5.
p) On the SQL Performance Analyzer Task Result page, you can clearly see that the
second trial is much faster than the original one. You should see that all five SQL
statements are improved in the second trial due a changed execution plan.
q) To get more details, analyze the differences in execution plan for all five
statements. You can do so directly from the SQL Performance Analyzer Task
Result page by clicking each SQL ID in the Top 10 table. Each time you click a
SQL ID, you can see in the SQL Details section all statistics differences between
the two trials as well as the differences in execution plans.

r) After this is done, go back to the SQL Performance Analyzer page (Home >
Software and Support > SQL Performance Analyzer), and delete
MY_SPA_TASK by selecting it and clicking Delete. On the Confirmation page,
click Delete.
s) Log out from Enterprise Manager.
6) From a terminal session, execute the revert.sh script to return to the situation you
were in before you started the lab.
[oracle@edrsr33p1-orcl SQL_Access_Advisor]$ ./revert.sh

2008

n s e
ice
b l el
Connected to:
fe r a
a n s
-tr
Production
n
o
an
SQL> SQL> SQL> SQL> SQL> SQL> SQL> SQL>ha

s eฺSQL> Connected.
) i d
SQL> SQL>
SQL> SQL>
Grant succeeded. i r ฺes nt Gu
g p ฺm ude
SQL> SQL> d S t
@ s
ia thi
User altered.
( j a b e
iaSQL>
SQL> SQL> Connected. u s
b
SQL> SQL> o
A 8 9t 10 11SQL>12SQL>13SQL>14SQL>15SQL>16 2 17 3 18 4 19 5
SQL> SQL>
6 io 7
n 21 22 23 24
to20
A n PL/SQL procedure successfully completed.
o s e
J SQL> SQL> SQL> SQL> SQL> SQL>
Table dropped.
SQL> SQL>
System altered.
SQL> SQL>
Table dropped.
SQL> SQL>
Table dropped.
SQL> SQL> SQL>

SQL>
SQL> SQL>

SQL> SQL> SQL>
SQL> SQL> SQL> SQL>

Materialized view log dropped.
SQL> SQL>
SQL> SQL>
SQL> SQL>
n s e
SQL> SQL>
i ce
Materialized view dropped.
b l el
SQL> SQL>
fer a
Index dropped.
a n s
o n -tr
an
SQL> SQL>
Table dropped.
h s
a deฺ
SQL> SQL>
e )
s Gui
Table dropped. r ฺ
i ent
p ฺ m d
d g S t u
SQL> SQL>
b
Revoke succeeded. i a@ this
i a (jaRem use
A b $Header:tosh_main.sql 06-mar-2008.15:00:45 cbauwens Exp $
SQL> SQL> SQL>
o
i Rem
SQL> Rem
n
ntoSQL>
e A SQL> Rem sh_main.sql
J o s SQL>
SQL>
Rem
Rem Copyright (c) 2001, 2008, Oracle. All rights reserved.
SQL> Rem
SQL> Rem NAME
SQL> Rem sh_main.sql - Main schema creation and load script
...
SQL> Rem
SQL>
SQL> SET ECHO OFF
specify directory path for the data files as parameter 5:
writeable directory path for the log files as parameter 6:

Session altered.
User dropped.
old 1: CREATE USER sh IDENTIFIED BY &pass

new 1: CREATE USER sh IDENTIFIED BY sh
User created.

n s e
i ce
b l el
User altered.
fer a
a n s
n -tr
o
an
User altered.
h s
a deฺ
e )
s Gui
r ฺ
Grant succeeded.
p ฺ mi dent
d g Stu
Grant succeeded.bia
@ his
j a e t
(
ia to u s
... A
b
o n io
nt
e A
J o s Grant succeeded.
Grant succeeded.
Connected.
Grant succeeded.
old 1: CREATE OR REPLACE DIRECTORY data_file_dir AS '&data_dir'

Directory created.
old 1: CREATE OR REPLACE DIRECTORY log_file_dir AS '&log_dir'

new 1: CREATE OR REPLACE DIRECTORY log_file_dir AS '/home/oracle/'

Directory created.
Grant succeeded.
Grant succeeded.
Grant succeeded.
Connected.
Session altered.
n s e
Session altered.
i ce
b l el
Table created.
fer a
a n s
n -tr
o
an
Table created.
h s
a deฺ
...
e )
s Gui
r ฺ
i ent
p ฺ m
Table created.
d g Stud
b i a@ this
(ja us...
Creating constraints
i a
e
b
altered. t
Table A
o
i o
nton
e A
J o s ...
Table altered.
specify path for data files as parameter 2:
specify path for log files as parameter 3:
Looking for indexes that could slow down load ...
no rows selected
loading TIMES using:

3.ctl
3.dat
/home/oracle/time_v3.log

2008
n s e
loading COUNTRIES using:
i ce
b l el
3.ctl
fer a
3.dat
a n s
/home/oracle/coun_v3.log
o n -tr
SQL*Loader: Release 11.1.0.6.0 - Production a on n Thu Mar 20 15:22:24
2008
h a s eฺ
) i d
i r ฺesAllnrights
t Gu reserved.
g p ฺm ude
t
d recordScount
i a @
Load completed - logical
h i s 23.
b
ja use t
ia ( tousing:
loading bCUSTOMERS
i oA
/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/cust_v
n
nto/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/cust1v
3.ctl
e A
J o s 3.dat
/home/oracle/cust1v3.log

2008

loading PRODUCTS using:

/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/prod_v
3.ctl

/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/prod1v
3.dat
/home/oracle/prod1v3.log

2008
loading PROMOTIONS using:

/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/prom_v
3.ctl
n s e
/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/prom1v
i ce
3.dat
b l el
/home/oracle/prom1v3.log
fer a
n s
a
2008
o n -tr
Copyright (c) 1982, 2007, Oracle. All rights n
a reserved.
h s eฺ
acountd10.
Save data point reached - logical record )
s count
erecord i
u 20.
Save data point reached - logical
i r ฺ t G
...
g p ฺm den count 500.
logicalurecord
d
Save data point reached -
S t
a @
Load completed -bilogical
s
hi count 503.
j a e trecord
b ia ( to us
i o A CHANNELS using:
loading
n
nto3.ctl
e A
J o s /u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/chan_v
3.dat
/home/oracle/chan_v3.log

2008
loading SALES using:

/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/sale_v
3.ctl
/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/sale1v
3.dat
/home/oracle/sale1v3.log

2008

...
loading COSTS using external table
Table created.
n s e
i ce
b l el
82112 rows created.
fer a
a n s
loading additonal SALES using:
o n -tr
an
v3.ctl
h s
a deฺ
v3.dat
e )
s Gui
/home/oracle/dmsal_v3.log r ฺ
i ent
p ฺ m d
d g
SQL*Loader: Release 11.1.0.6.0
S
- u
t Production on Thu Mar 20 15:22:38
2008
b i a@ this
Copyright (c)
i a (ja1982,u2007,
se Oracle. All rights reserved.
A o
b point treached
i o
Save data
n
- logical record count 100.
ntoSave data point reached - logical record count 2800.

...
e A
J o s Load completed - logical record count 2804.
loading SUPPLEMENTARY DEMOGRAPHICS using:

/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/dem_v3
.ctl
/u01/app/oracle/product/11.1.0/db_1/demo/schema/sales_history/dem1v3
.dat
/home/oracle/dem1v3.log

2008

...

Commit complete.
Enabling constraints ...

Table altered.
...
Table altered.
n s e
Creating additional indexes ...
i ce
b l el
Index created.
fer a
a n s
...
n - tr
o
s an
Index created.
) h a deฺ
r ฺ e s Gui
Create dimensions ...
p ฺ mi dent
Dimension created. @d
g Stu
b i a t h is
i a (ja use
b
Commit complete.
A to
o
ni procedure successfully completed.
ntoPL/SQL
e A
J o s
no rows selected
Dimension created.
no rows selected
Dimension created.
no rows selected

Dimension created.

no rows selected
Dimension created.
n s e
i ce
no rows selected
b l el
Creating MVs as tables ...
fer a
a n s
n -tr
o
an
View created.
h s
a deฺ
Table created.
e )
s Gui
r ฺ
i ent
p ฺ m
Table created.
d g Stud
b i a@ this
Index created.
i a (ja use
A b to
o
ni created.
n t oIndex
e A
Jos Index created.
Index created.
Creating materialized views ...
Creating comments ...
Comment created.
...

Comment created.

-
<<<<< CREATE CATALOG sh_cat for Sales History >>>>>
Catalog Dropped
CWM Collect Garbage
-
<<<<< CREATE the Sales CUBE >>>>>
Sales amount, Sales quantity
<TIMES CHANNELS PRODUCTS CUSTOMERS PROMOTIONS >
Drop SALES_CUBE prior to recreation
n s e
Cube Dropped
i ce
Add dimensions -
b l el
to SALES_CUBE and map the foreign keys
Create measures -
fer a
for SALES_CUBE and map to columns in the fact table
a n s
n - tr
o
an
Add SALES_CUBE to the catalog
h s
a deฺ
SALES_CUBE successfully added to sh_cat
-
e )
s Gui
r ฺ
mi dent
<<<<< CREATE the Cost CUBE >>>>>
p ฺ
Unit Cost, Unit Price < TIMES PRODUCTS CHANNELS PROMOTIONS >
g Stu
d
Drop COST_CUBE prior to recreation
Cube Dropped
b i a@ this
Add dimensions -
i a (ja use
to COST_CUBE and map the foreign keys
b to
Create measures -
A
o n io for COST_CUBE and map to columns in the fact table
nt
e A to SUM for all measures over TIME
J o s Add COST_CUBE to the catalog

COST_CUBE successfully added to sh_cat
-
<<<<< TIME DIMENSION >>>>>
- Period Number of Days created
- Period End Date created
- Time dimension
- Long description
- Day name

- Short Description
- Day name

- Time Span
- Days in calendar month
- Days in calendar quarter
- Days in calendar year
- Days in fiscal month
- Days in fiscal quarter
- Days in fiscal year
n s e
- End Date
i ce
- End of calendar month
b l el
-
-
End of calendar quarter
End of calendar year
fer a
- End of fiscal month
a n s
- End of fiscal quarter
o n -tr
an
- End of fiscal year
-
<<<<< CUSTOMERS DIMENSION >>>>>
h s
a plural ฺ
ename
Dimension - display name, description) and i d
i r ฺ es t Gu
Hierarchy - display name andฺm
g p description
u d en
d hierarchy
- default calculation
t
hierarchy
S
@
- default display
i s
Level Attributesbi-aname, tdisplay
hprior name, description
Drop dimension j a
(to dropus e
attributes to re-creation
i a
b to tdrop
No attribute o
o A
i attribute to drop
No attribute
n
toNo
A n No attribute to drop
o s e No attribute to drop
J No attribute to drop
- Other Customer Information created
- Long Description
- Short Description
-
<<<<< PRODUCTS DIMENSION >>>>>


- Long Description
- Short Description
-
<<<<< PROMOTIONS DIMENSION >>>>>
n s e
i ce
b l el
fer a
a n s
n - tr
o
an
- Long Description
h s
a deฺ
- Short Description
e )
s Gui
r ฺ
mi dent
-
p ฺ
g Stu
d
b i a@ this
i a (ja use
b to
A
o n io
nt
e A No attribute to drop
J o s Create dimension attributes and add their level attributes

- Long Description
- Short Description
-
Commit complete.


--------------------------------------------------------------
#!/bin/bash
n s e
i ce
b l el
export
fer a
n s
a
n -tr
cp * $ORACLE_HOME/demo/schema/sales_history a n
o
h a s eฺ
e s ) u id
ฺ
ir ent G
SET ECHO ON ฺ m
p tud
SET FEEDBACK 1
d g
SET NUMWIDTH 10
i a @ his S
SET LINESIZE 8000
SET TRIMSPOOL ( j aONb se t
ia to u
SET TAB bOFF
i A
oLONG
SET PAGESIZE 100
to n
SET 1000
A n
o s e CONNECT / AS SYSDBA
J grant dba to sh;
connect sh/sh
set serveroutput on size 32768;

set echo on;
variable norecs number;
Rem Clean up
declare
name varchar2(30);
cursor name_cur1 is
select task_name from user_advisor_templates
where task_name like '%SQLACCESS%';

begin
----------------------------------------------------------------
-----------
-- Get rid of templates, tasks and workloads.
----------------------------------------------------------------
-----------
open name_cur1;
loop
fetch name_cur1 into name;
exit when name_cur1%NOTFOUND;
dbms_advisor.update_task_attributes(name,null,null,'FALSE','FALSE');
dbms_advisor.delete_task(name);
n s e
end loop;
i ce
b l el
close name_cur1;
fer a
end;
a n s
/
n - tr
o
s an
Rem make a temp table
) h a deฺ
r ฺ e s Gui
DROP TABLE temp_table purge;ฺmi e nt
d p
g Stu d
b i a@ this
(ja upurge;
drop table tempjfv
i a se
A to
b customersjfv
i o
drop table
n
purge;
nto
e A
J o s execute dbms_advisor.delete_task('%');
execute dbms_advisor.delete_sqlwkld('%');
execute dbms_sqltune.drop_sqlset('SQLSET_MY_SQLACCESS_WORKLOAD');
EXECUTE DBMS_STATS.UNLOCK_SCHEMA_STATS('SH');
DROP MATERIALIZED VIEW LOG ON "SH"."CUSTOMERS";
DROP MATERIALIZED VIEW LOG ON "SH"."CHANNELS";
DROP MATERIALIZED VIEW LOG ON "SH"."TIMES";
DROP MATERIALIZED VIEW LOG ON "SH"."SALES";
DROP MATERIALIZED VIEW "SH"."MV_01DF0000";
DROP INDEX "SH"."CUSTOMERS_IDX_01DF0002";

DROP TABLE "SH"."CUSTOMERS" PURGE;
DROP TABLE "SH"."CUSTOMERS11" CASCADE CONSTRAINTS PURGE;
connect / as sysdba
revoke dba from sh;
@sh_main sh example temp oracle

/home/oracle/ v3
exit;
FIN!
n s e
i ce
b l el
fer a
a n s
n - tr
o
s an
) h a deฺ
r ฺ e s Gui
p ฺ mi dent
d g Stu
b i a@ this
i a (ja use
Ab to
o n io
nt
e A
J o s

Practice 11-3: Using Automatic SQL Tuning
In this practice, you manually launch Automatic SQL Tuning to automatically tune a
small application workload. You then investigate the outcomes and configuration
possibilities.
1) On the Server page, click Automated Maintenance Tasks, check that Status is set to
Enabled, and click Configure. Click the Configure button next to Automatic SQL
Tuning. Select Yes for “Automatic Implementation of SQL Profiles.” Then, click
Apply. Execute the ast_setup.sh script from a terminal window connected as the
oracle user. This script creates the AST user used throughout this practice, turns off
automatic maintenance tasks, and drops any existing profiles on queries executed by
the AST user.
n s e
$ cd /home/oracle/solutions/Automatic_SQL_Tuning
ice
b l el
$ ./ast_setup.sh
f er a
an s
[oracle@edrsr33p1-orcl Automatic_SQL_Tuning]$ ./ast_setup.sh
- t r
n 18 15:31:49
SQL*Plus: Release 11.1.0.6.0 - Production on Tue n oMar
2008
s a
Copyright (c) 1982, 2007, Oracle. s ) a eฺ
hrights idreserved.
i r ฺe nt Gu
All
g p ฺm ude
Connected to:
Oracle Database 11g @
d S t
b i a t h i
Enterprises Edition Release 11.1.0.6.0 -
(ja useOracle Label Security, OLAP, Data Mining

Production
i a
to Testing options
b Application
A
and Real
o
ni SQL> SQL> SQL> drop user ast cascade
toSQL>
A n *
o s e ERROR at line 1:
J ORA-01918: user 'AST' does not exist
SQL> SQL>
User created.
SQL> SQL>
Grant succeeded.
SQL> SQL>
System altered.

System altered.

SQL> SQL> SQL> SQL> SQL> SQL> 2 3 4 5 6 7 8 9


Practice 11-3: Using Automatic SQL Tuning (continued)
[oracle@edrsr33p1-orcl Automatic_SQL_Tuning]$
--------------------------------------------------------------
#!/bin/bash
cd /home/oracle/solutions/Automatic_SQL_Tuning
n s e
i ce
export
b l el
fer a
a n s
n -tr
o
set echo on
s an
) h a deฺ
drop user ast cascade;
r ฺ e s Gui
create user ast identified by ฺ i ent
mast;
d g p t u d
grant dba to ast; @
a is S
b i t h
a shared_pool;
alter system(jflush
i a u se
A b to
--
n o
i Turn off AUTOTASK
nto--
e A --
J o s alter system set "_enable_automatic_maintenance"=0;
--
-- Clear out old executions of auto-sqltune
--
exec dbms_sqltune.reset_tuning_task('SYS_AUTO_SQL_TUNING_TASK');
--
-- Drop any profiles on AST queries
--
declare
cursor prof_names is
select name from dba_sql_profiles where sql_text like '%AST%';
begin
for prof_rec in prof_names loop
dbms_sqltune.drop_sql_profile(prof_rec.name);
end loop;
end;

/
FIN!
2) In preparation for the practice, you should execute a workload. Execute the
run_workload_stream.sh script. This script executes, multiple times, a query that
is not correctly optimized. The query in question uses hints that force the optimizer to
pick a suboptimal execution plan. The script execute for approximately 30 seconds.
./run_workload_stream.sh
Tue Mar 18 15:38:05 GMT-7 2008
n s e
ice
el
2008
a b l
fe r
a n s
o n -tr
Connected to:
s a n11.1.0.6.0 -
Oracle Database 11g Enterprise Edition Release
Production
)
With the Partitioning, Oracle LabelsSecurity,
eฺ Data Mining
ha idOLAP,
i r ฺe nt Gu
g p ฺm ude
SQL> SQL> SQL> SQL> d S t
no rows selected a@ s
j a b i
e thi
SQL>
ia ( to us
no rows bselected
n i oA
ntoSQL>
e A no rows selected
J o s ...
SQL>
no rows selected
SQL>
no rows selected
SQL>
no rows selected
SQL>
no rows selected
SQL>
no rows selected


Tue Mar 18 15:38:23 GMT-7 2008
--------------------------------------------------------------
#!/bin/bash
export
n s e
i ce
b l el
date
fer a
sqlplus ast/ast <<FIN!
a n s
o n -tr
set echo on
n
a c.cust_id,
select /*+ USE_NL(s c) FULL(s) FULL(c) AST
h s
a deฺ c where s.cust_id
*/
sum(s.quantity_sold) from sh.sales s,) sh.customers
r ฺ e G ui
sby c.cust_id;
= c.cust_id and c.cust_id < 2 group
ฺ m i ent
…
d p
g Stud
select /*+ USE_NL(sb i a@ t h is FULL(c) AST */ c.cust_id,
c) FULL(s)
ja usfrom
(and
sum(s.quantity_sold)
a
e sh.sales s, sh.customers c where s.cust_id
b i
= c.cust_id
to
c.cust_id < 2 group by c.cust_id;
io A
t o n
FIN!
A n
o s e date
J 3) Automatic SQL Tuning is implemented using an automated task that runs during
maintenance windows. However, you do not wait for the next maintenance window to
open. Instead, you force the opening of your next maintenance window now. This
automatically triggers the Automatic SQL Tuning task. Execute the run_ast.sh
script to open your next maintenance window now. The script’s execution takes a
couple of minutes.
./run_ast.sh
[oracle@edrsr33p1-orcl Automatic_SQL_Tuning]$ ./run_ast.sh

Tue Mar 18 15:43:48 GMT-7 2008

2008

Connected to:
Production
SQL> SQL> SQL> SQL>

SQL> SQL> SQL> SQL> 2 3 4

SQL> SQL>
WINDOW
--------------------------------------------------------------------
n s e
------------
i ce
TUESDAY_WINDOW
b l el
SQL> SQL> SQL> SQL> SQL> SQL> 2
fer a
System altered.
a n s
o n -tr
an
SQL> SQL> >
h s
a deฺ
SQL> SQL> >
e )
s Gui
r ฺ
i ent
p ฺ m d
SQL> SQL>
d g S t u
b i a@ this completed.
SQL> SQL> SQL>

i a (jaSQL> u se SQL> completed.
SQL>
b
PL/SQL procedure
A tosuccessfully
n o
i SQL> 2 3 4 5 6 7 8 9 10
toSQL> 11 12 13
e An 14 15 16 17 18 19 20 21 22
Jos
SQL> SQL> 2
System altered.
SQL> SQL> SQL> SQL> SQL> SQL> SQL> >

SQL> SQL> >

SQL> SQL> SQL> Disconnected from Oracle Database 11g Enterprise

Tue Mar 18 15:44:50 GMT-7 2008
--------------------------------------------------------------
#!/bin/bash

export
date
set echo on
n s e
i ce
b l el
variable window varchar2(20);
fer a
a n s
begin
o n -tr
an
select upper(to_char(sysdate,'fmday'))||'_WINDOW' into :window from
dual;
end;
h s
a deฺ
/
e )
s Gui
r ฺ
print window;
p ฺ mi dent
d g Stu
--
b i a@ thmaintenance
-- Open the corresponding is window, but with other clients
j a e
ia ( to us
disabled
-- b
n i o Asystem set "_enable_automatic_maintenance"=1
nto/
alter
e A
J o s exec dbms_auto_task_admin.disable( -
'auto optimizer stats collection', null, :window);
exec dbms_scheduler.open_window(:window, null, true);
--
-- Close the maintenance window when sqltune is done
--
declare
running number;
begin
loop
select count(*)

into running
from dba_advisor_executions
where task_name = 'SYS_AUTO_SQL_TUNING_TASK' and
status = 'EXECUTING';
if (running = 0) then
exit;
end if;
dbms_lock.sleep(60);
end loop;
dbms_scheduler.close_window(:window);
end;
/
n s e
ice
b l el
/
fe r a
--
a n s
-
-- Re-enable the other guys so they look like they are enabled in
n tr
o
an
EM.
-- Still they will be disabled because we have set the underscore.
--
h s
a deฺ
e )
s Gui
r ฺ
mi denull, nt :window);
p ฺ
'auto optimizer stats collection',
g Stu-
d
b a@ null,
'auto space advisor', t h is :window);
exec dbms_auto_task_admin.enable(
i
i a (ja use
FIN! A
b to
i o
nton
e A date
J o s 4) Execute the run_workload_stream.sh script again. What do you observe?

a) You should see that the execution time for run_workload_stream.sh is much
faster than the original execution. This is probably due to the fact that Automatic
SQL Tuning implemented a profile for your statement automatically.
Tue Mar 18 15:46:42 GMT-7 2008

2008
Connected to:

Production
SQL> SQL> SQL> SQL>

no rows selected
SQL>
no rows selected
SQL>
no rows selected
...
n s e
SQL>
i ce
no rows selected
b l el
SQL>
fer a
no rows selected
a n s
o n -tr
an
SQL>
no rows selected
h s ฺ
a 11gdeEnterprise
)
s Gu
SQL> SQL> Disconnected from Oracle Database
Release 11.1.0.6.0 - Production rฺe
i Edition
With the Partitioning, Oracle ฺ m i Security,

Label e nt OLAP, Data Mining
g p
d 2008St
and Real Application Testing u d
options
Tue Mar 18 15:46:42 @
b i
GMT-7
[oracle@edrsr33p1-orcl t h is
a Automatic_SQL_Tuning]$
i a (ja use
b to
------------------------------------------------------------
A
o
ni /home/oracle/solutions/AST
#!/bin/bash
ntocd
e A
export
l/bin:/usr/X11R6/bin
date
sqlplus ast/ast <<FIN!
set echo on
select /*+ USE_NL(s c) FULL(s) FULL(c) AST */ c.cust_id,

sum(s.quantity_sold) from sh.sales s, sh.customers c where s.cust_id
= c.cust_id and c.cust_id < 2 group by c.cust_id;

select /*+ USE_NL(s c) FULL(s) FULL(c) AST */ c.cust_id,
sum(s.quantity_sold) from sh.sales s, sh.customers c where s.cust_id
= c.cust_id and c.cust_id < 2 group by c.cust_id;
FIN!
date
5) Force the creation of an Automatic Workload Repository (AWR) snapshot.
./create_snapshot.sh

2008
n s e
ce
eli
a b l
Connected to:
fe r
a n s
Production
o n -tr
s an
) h a deฺ
SQL> SQL> SQL> SQL>
r
PL/SQL procedure successfully completed.ฺ e s Gui
p ฺ mi Database
Oracle d e nt 11g Enterprise Edition
SQL> SQL> Disconnected from
d g Stu
a@ Oracle
b i
t h is Label Security, OLAP, Data Mining
( j a
s eAutomatic_SQL_Tuning]$
Testing options
bia to
[oracle@edrsr33p1-orcl u
io A
t o n
--------------------------------------------------------------
A n
o s e #!/bin/bash
J cd /home/oracle/solutions/Automatic_SQL_Tuning
export
set echo on
FIN!

6) How would you confirm that a SQL Profile was automatically implemented?
a) In Oracle Enterprise Manager, locate the Automatic SQL Tuning summary page
under Server > Automated Maintenance Tasks > Automatic SQL Tuning. The
task has already run in one maintenance window and has results ready to be
viewed.
b) View the tuning results.

c) Look at the graphs on the Task Activity Summary page.
d) Focus on understanding the pie chart and the bar graph next to it. You should be
able to get a feeling for the general finding breakdown, as well as the number of
SQL profiles implemented by the task.
e) Click View Report to see a detailed SQL-level report. Find the SQL that ran in the
n s e
AST schema. Note the green check mark meaning that the profile was
i ce
implemented.
b l el
f) Click the corresponding option button and then View Recommendations. fe r a
a n s
o n -tr
g) Click the Compare Explain Plans eyeglass icon for the SQL Profile entry.
h) View the old and new explain plans for the query. a n
i) Then click the “Recommendations for SQL_ID” h a slocatorelink
ฺ to return to the
) i d
previous screen.
i r ฺes nt Gu
j) Investigate a SQL profile. While
g p ฺmstill uondthe
e “Recommendations for SQL_ID”
d t
i a @ his S SQL Details page for this SQL.
page, click the SQL text to go to the
k) This takes you j e t History tab. Note the link to
atob the Tuning
(
ia to u
SYS_AUTO_SQL_TUNING_TASK s that is there to show that the SQL was tuned
b
n i obyAthis tuning task.
Antol) Look at the Plan Control subpage and note that a profile was created
e automatically for this SQL. The AUTO type means it was automatically created.
Jos m) Click the Statistics tab to take a look at the execution history for this SQL.
n) Depending on the speed of your machine, you may not see two hash values. If
that is the case, ignore this step and the following one. Select Real Time: Manual
Refresh from the View Data and then each of possible two Plan Hash Values from
the corresponding drop-down list. Choose one after the other and wait for the
page to refresh each time.
o) Depending on the speed of your environment, you should see one statement with
a relatively high elapsed time per execution, and one with very low elapsed time
per execution. This shows the improved plan. If you select All from the Plan Hash
Values drop-down list, you might not be able to see the execution corresponding
to the statement after tuning in the Summary graph. This might be because the
workload was too short to execute.
7) Generate a text report for more indepth information. From the command line, execute
the get_task_report.sh script. What do you observe?

a) Note the first queries that fetch execution name and object number from the
advisor schema, followed by the final query that gets the text report. In the text
report, look for the section about the SQL profile finding and peruse the
Validation Results section. This shows you the execution statistics observed
during test-execute and allows you to get a better idea about the profile’s quality.
You can also use the report_auto_tuning_task API to get reports that span
multiple executions of the task.
./get_task_report.sh
[oracle@edrsr33p1-orcl Automatic_SQL_Tuning]$ ./get_task_report.sh

2008
n s e
i ce
b l el
fe r a
a n s
-tr
Connected to:
n
o
an
Production
and Real Application Testing options h s

a deฺ
e )
sSQL> SQL>ui
r ฺ G
Session altered. p ฺ mi dent
SQL> SQL> SQL> SQL> SQL> SQL> SQL>
d g Stu
3a@4 is STATUS
EXECUTION_NAMEabi
SQL> SQL> 2
j e t h EXECUTION_START
(
a to u s
EXEC_1 bi
------------------------------ ----------- -------------------
o A COMPLETED 03/18/2008 15:43:54

ni SQL> SQL> SQL> 2 3 4 5 6 7
toSQL>
A n PL/SQL procedure successfully completed.
o s e
J SQL> SQL>
LAST_EXEC
--------------------------------------------------------------------
------------
EXEC_1
SQL> SQL> SQL> SQL> SQL> SQL> SQL> SQL> 2 3 4 5 6 7

8 9 10
SQL> SQL>
OBJ_ID
----------
3
SQL> SQL> SQL> SQL> SQL> SQL> 2 3 GENERAL INFORMATION SECTION

--------------------------------------------------------------------
-----------
Tuning Task Name : SYS_AUTO_SQL_TUNING_TASK
Tuning Task Owner : SYS

Workload Type : Automatic High-Load SQL
Workload
Scope : COMPREHENSIVE
Global Time Limit(seconds) : 3600
Per-SQL Time Limit(seconds) : 1200
Completion Status : COMPLETED
Started at : 03/18/2008 15:43:54

Completed at : 03/18/2008 15:44:14
Number of Candidate SQLs : 4
Cumulative Elapsed Time of SQL (s) : 27
--------------------------------------------------------------------
-----------
Object ID : 3
Schema Name: AST
SQL ID : by9m5m597zh19
n s e
SQL Text : select /*+ USE_NL(s c) FULL(s) FULL(c) AST */
i ce
c.cust_id,
b l el
where
sum(s.quantity_sold) from sh.sales s, sh.customers c
fe ra
n
s.cust_id = c.cust_id and c.cust_id < 2 group by
a s
c.cust_id
o n -tr
an
--------------------------------------------------------------------
s
-----------
) h a deฺ
FINDINGS SECTION (2 findings)
r ฺ e s Gui
mi dent
--------------------------------------------------------------------
-----------
p ฺ
g explain
d S tu plans section below)
1- SQL Profile Finding
b i a@ this
(see
--------------------------------------------------------
i a (ja"SYS_SQLPROF_01463043cc730000"
A potentially better
u seexecution plan was found for this statement.
A b
SQL profile
to was created
o statement.
ithis
automatically
n
for
nt o
e A
J o s Recommendation (estimated benefit: 98.62%)
------------------------------------------
- An automatically-created SQL profile is present on the system.
Name: SYS_SQLPROF_01463043cc730000
Status: ENABLED
Validation results
------------------
The SQL profile was tested by executing both its plan and the
original plan
and measuring their respective execution statistics. A plan may
have been
only partially executed if the other could be run to completion in
less time.
Original Plan With SQL Profile %

Improved
------------- ---------------- --------
--
Completion Status: COMPLETE COMPLETE

Elapsed Time(ms): 182 0
100%
CPU Time(ms): 182 0
100%
User I/O Time(ms): 0 0
Buffer Gets: 3177 44
98.61%
Disk Reads: 0 0
Direct Writes: 0 0
Rows Processed: 0 0
Fetches: 0 0
Executions: 1 1
2- Index Finding (see explain plans section below)

--------------------------------------------------
The execution plan of this statement can be improved by creating
n s e
one or more
i ce
indices.
b l el
Recommendation (estimated benefit: 90.97%)
fer a
------------------------------------------
a n s
o n -tr
- Consider running the Access Advisor to improve the physical
an
schema design
or creating the recommended index.
h s
a deฺ
create index SH.IDX$$_00010001 on SH.SALES("CUST_ID");
e )
s Gui
Rationale r ฺ
i ent
---------
p ฺ m d significantly improves the
d g
Creating the recommended
t u
indices
S
execution plan
b i a@ However,
of this statement. t h is it might be preferable to run
"Access Advisor"
a ja use
(representative
usingi
b This to
a SQL workload as opposed to a single
A
iowill allow to get comprehensive index recommendations which
statement.
o n
t takes into
A n
o s e account index maintenance overhead and additional space
J consumption.
--------------------------------------------------------------------
-----------
EXPLAIN PLANS SECTION
--------------------------------------------------------------------
-----------
1- Original With Adjusted Cost

------------------------------
--------------------------------------------------------------------
------------
-------------------
(%CPU)| Time
| Pstart| Pstop |
--------------------------------------------------------------------
------------

-------------------
(2)| 00:00:1
1 | | |
| 1 | HASH GROUP BY | | 1 | 13 | 902
(2)| 00:00:1
1 | | |
| 2 | NESTED LOOPS | | 1 | 13 | 901
(2)| 00:00:1
1 | | |
|* 3 | TABLE ACCESS FULL | CUSTOMERS | 1 | 5 | 405
(1)| 00:00:0
5 | | |
| 4 | PARTITION RANGE ALL| | 1 | 8 | 495
(3)| 00:00:0
6 | 1 | 28 |
n s e
|* 5 | TABLE ACCESS FULL | SALES | 1 | 8 | 495
i ce
(3)| 00:00:0
b l el
6 | 1 | 28 |
fer a
--------------------------------------------------------------------
------------
a n s
-------------------
o n -tr
s a nid):
h a deฺ
---------------------------------------------------
)
3 - filter("C"."CUST_ID"<2) rฺe
s Gui
5 - filter("S"."CUST_ID"<2 ฺ i "S"."CUST_ID"="C"."CUST_ID")
mAND e nt
p
g Stu d
2- Using SQL Profile@d
b i
-------------------- a t h is
i a (ja 3070788227
Plan hash value:
u se
A b to
i o
--------------------------------------------------------------------
n
nto------------
e A ---------------------------------------
J o s | Id | Operation
| Bytes |
| Name | Rows
Cost (%CPU)| Time | Pstart| Pstop |

--------------------------------------------------------------------
------------
---------------------------------------
1 | 13 |
55 (2)| 00:00:01 | | |
| 1 | HASH GROUP BY | |
1 | 13 |
55 (2)| 00:00:01 | | |
1 | 13 |
54 (0)| 00:00:01 | | |
| 3 | PARTITION RANGE ALL | |
1 | 8 |
54 (0)| 00:00:01 | 1 | 28 |
| 4 | TABLE ACCESS BY LOCAL INDEX ROWID| SALES |
1 | 8 |
54 (0)| 00:00:01 | 1 | 28 |

| 5 | BITMAP CONVERSION TO ROWIDS | |
| |
| | | |
|* 6 | BITMAP INDEX RANGE SCAN | SALES_CUST_BIX |
| |
| | 1 | 28 |
|* 7 | INDEX UNIQUE SCAN | CUSTOMERS_PK |

1 | 5 |
0 (0)| 00:00:01 | | |
--------------------------------------------------------------------
------------
---------------------------------------

---------------------------------------------------
n s e
6 - access("S"."CUST_ID"<2)
i ce
filter("S"."CUST_ID"<2)
b l el
filter("C"."CUST_ID"<2)
fer a
a n s
3- Using New Indices
o n -tr
an
--------------------
h s
a deฺ
)
s Gui
--------------------------------------------------------------------
e
------------ r ฺ
i ent
ฺ m
---------------------------------------
p d
| Id | Operation
d g S t u | Name | Rows
| Bytes |
Cost (%CPU)| Time b i a@| Pstart|
t h is Pstop |
i a (ja use
--------------------------------------------------------------------
b
------------
A to
i o
---------------------------------------
nto|1n | 0 |13SELECT STATEMENT | |
e A |
J o s |
5 (0)| 00:00:01 |
1 | SORT GROUP BY NOSORT
| |
| |
1 | 13 |
5 (0)| 00:00:01 | | |
| |
| | | |
1 | 13 |
5 (0)| 00:00:01 | | |
|* 4 | INDEX RANGE SCAN | CUSTOMERS_PK |
1 | 5 |
2 (0)| 00:00:01 | | |
|* 5 | INDEX RANGE SCAN | IDX$$_00010001 |
1 | |
2 (0)| 00:00:01 | | |
| 6 | TABLE ACCESS BY GLOBAL INDEX ROWID| SALES |
1 | 8 |
3 (0)| 00:00:01 | ROWID | ROWID |
--------------------------------------------------------------------
------------

---------------------------------------

---------------------------------------------------
4 - access("C"."CUST_ID"<2)
filter("S"."CUST_ID"<2)
--------------------------------------------------------------------
-----------

n s e
i ce
b l el
fe
-------------------------------------------------------------- r a
a n s
#!/bin/bash
o n -tr
s an
) h a deฺ
r ฺ e s Gui
ฺ mi dent
p
d g Stu
export
b i a@ this
i a (ja use
A to <<FIN!
b/ as sysdba
o
ni echo on
sqlplus
ntoset
e A
J o s set long 1000000000
--
-- Check the execution names
--
alter session set nls_date_format = 'MM/DD/YYYY HH24:MI:SS';
select execution_name, status, execution_start

where task_name = 'SYS_AUTO_SQL_TUNING_TASK'
order by execution_start;
variable last_exec varchar2(30);
begin
select max(execution_name) keep (dense_rank last order by
execution_start)
into :last_exec
where task_name = 'SYS_AUTO_SQL_TUNING_TASK';

end;
/
print :last_exec
--
-- Find the object ID for query AST with sql_id by9m5m597zh19

--
variable obj_id number;
begin
select object_id
into :obj_id
from dba_advisor_objects
n s e
execution_name = :last_exec and
i ce
type = 'SQL' and
b l el
end;
attr1 = 'by9m5m597zh19';
fe r a
/
a n s
o n -tr
an
print :obj_id
--
h s eฺ
a onedquery
-- Get a text report to drill down on) this i
--
i r ฺ es t Gu
set pagesize 0
g p ฺm uden
d 'TEXT', t 'TYPICAL', 'ALL', :obj_id)
select dbms_sqltune.report_auto_tuning_task(
S
@ s
thi
:last_exec, :last_exec,
from dual;
a b ia
j e
FIN! b ia ( to us
n i o A how to configure Automatic SQL Tuning using Enterprise Manager.
8) Investigate
to
e An a) Back in EM, go to the Automated Maintenance Tasks page.
Jos b) The chart here shows times in the past when each client was executed, and times
in the future when they are scheduled to run again.
c) Modify the graph’s begin and end points with the widgets at the upper right.
d) Click the Configure button.
e) This brings you to the Automated Maintenance Tasks Configuration page.
f) From this page, you can disable individual clients and change which windows
they run in.
g) Disable the Automatic SQL Tuning client entirely, click Apply, and then click the
locator link to return to the last page.
h) Note that no light blue bars appear for Automatic SQL Tuning in the future.
i) Return to the configuration page, enable the task again, and click Apply to undo
your changes.

j) Click the Automatic SQL Tuning link on the Automated Maintenance Tasks
Configuration page.
k) This takes you to the page where you can configure the task itself, and set beyond
when it will run.
l) Note that there are more fine-grained controls here, such as one that allows the
task to run but not implement profiles, and one that allows you to control the
maximum number of profiles created per run.
9) Investigate how to configure Automatic SQL Tuning using PL/SQL. From your
terminal session, execute the manual_config.sh script. What does it do?
a) Note the first action. You changed the total time limit for the task. Instead of
running for an unlimited amount of time (still bound by the maintenance window
boundaries), it now runs for a maximum of one hour. The n s e
i ce
execute_tuning_task API call runs the task immediately, in the foreground.
Use this to run the task yourself whenever you want. b l el
fe r a
a n s
./manual_config.sh
o n -tr
an
--------------------------------------------------------------
s
) h a deฺ
#!/bin/bash
r ฺ e s Gui
p ฺ mi dent
d g Stu
b i a@ this
i a (ja use
A b to
ni o
export
ntoPATH=/u01/app/oracle/product/11.1.0/db_1/bin:/bin:/usr/bin:/usr/loca
e A l/bin:/usr/X11R6/bin:/usr/java/jdk1.5.0_11/bin
J o s sqlplus / as sysdba <<FIN!
connect / as sysdba
set echo on
--
-- Configure the task to run for at most 30 minutes. The value of
the
-- TIME_LIMIT parameter determines the total time allowed for a task
execution.
--
select parameter_value
from dba_advisor_parameters
parameter_name = 'TIME_LIMIT';
exec dbms_sqltune.set_tuning_task_parameter( -
'SYS_AUTO_SQL_TUNING_TASK', 'TIME_LIMIT', 1800);

select parameter_value
from dba_advisor_parameters
parameter_name = 'TIME_LIMIT';
--
-- Run the task immediately

--
exec dbms_sqltune.execute_tuning_task('SYS_AUTO_SQL_TUNING_TASK');
FIN!
10) Note: In your case, the task executes quickly because the workload to take into
account is really small. However, you could use the interrupt_task.sh script
from another session to stop the task, should it last too long. n s e
i ce
[oracle@edrsr33p1-orcl Automatic_SQL_Tuning]$ cat interrupt_task.sh
b l el
#!/bin/bash
fe r a
a n s
-tr
o n
an
s
a deฺ
h
e )
s Gui
export
r ฺ
ฺ mi dent
p
g Stu
d
sqlplus / as sysdba
b i a@
<<FIN! is
th
a
connect / as(jsysdbause
A biona to
setoecho
ni
to--
A n
o s e -- Interrupt the task
J --
exec dbms_sqltune.interrupt_tuning_task('SYS_AUTO_SQL_TUNING_TASK');
FIN!
11) Ensure that you disable automatic implementation of SQL profiles to clean up your
environment.
a) On the EM Server page, click Automated Maintenance Tasks.

b) Check that Status is set to Enabled, and click Configure.
c) Click the Configure button next to Automatic SQL Tuning.
d) Select No for “Automatic Implementation of SQL Profiles.”
e) Then, click Apply.

J
o s e
n A
t o n
Aio
b i a
b i
to
(ja use
d p ฺ
a@ this
r ฺ
g Stu
e ) h
mi dent
s
s Gui
an
o
a deฺ
tr n -
a n s fe
r
a b l el
i
ce n s e
J
o s e
n A
t o n
Aio
b i a
b i
to
(ja use
d p ฺ
a@ this
r ฺ
g Stu
e ) h
mi dent
s
s Gui
an
o
a deฺ
tr n -
a n s fe
r
a b l
Index
i el
ce n s e
A
Access path 3-16, 3-17, 3-19, 3-23, 3-24, 3-28, 3-32, 4-2, 4-3,
4-4, 4-6, 4-8, 4-10, 4-14, 4-17, 4-29, 4-31, 4-33, 4-43, 4-47,
4-56, 4-71, 4-72, 5-49, 6-8, 6-14, 9-2, 9-9, 9-10, 9-11, 9-12,
9-13, 9-14, 9-17, 9-30, 11-5, 11-7, 11-10, 11-15
Action 1-11, 1-12, 1-22, 1-23, 2-7, 2-9, 2-13, 2-14, 2-15, 2-17,
2-18, 2-19, 3-3, 3-12, 3-25, 3-30, 4-3, 4-6, 4-30, 4-70, 6-14,
7-20, 7-22, 7-32, 8-8, 10-4, 10-8, 10-9, 10-10, 10-13, 10-14, 11-3,
11-15, 11-21, 11-24, 11-25
Adaptive Cursor Sharing 8-2, 8-12, 8-13, 8-14, 8-15, 8-16, 8-17,
8-18, 8-19, 8-20, 8-21 n s e
i ce
ADDM 2-6, 2-8, 2-9, 2-10, 2-11, 2-14, 2-15, 5-27, 5-28, 11-4,
b l el
11-9, 11-12, 11-14, 11-26, 11-38 fer a
a n s
ADR 10-5, 10-14, 10-15
o n -tr
AGGREGATE 1-6, 1-17, 1-18, 2-19, 3-5, 3-20, 3-21, 3-29,n3-30,
4-20, 4-27, 4-30, 4-31, 4-48, 4-52, 4-53, 5-26, a s a5-47,
5-38, eฺ 6-10,
)h id
10-19, 10-25, 10-28, 10-29
i r ฺes nt Gu
ALL_ROWS 3-32, 3-33, 5-14, 9-4, 9-8,
g p ฺm9-10,ud10-26,
e 10-28, 10-29,
d t
11-7 @
ia
b 4-12, sS
hi
t4-20, 4-24, 4-27, 4-37, 4-38,
ALTER 1-8, 3-3, 3-11,
( j a3-28, s e
4-50, 5-11,b ia 5-44,
5-25, t o u7-8, 7-22, 7-26, 8-11, 8-12, 9-10, 10-12,
n
11-11 i oA
n to
A
eALTER SESSION 3-3, 5-44, 8-11, 8-12, 9-10, 10-12
Jos ALTER SYSTEM 3-3, 7-22, 8-11
AND-EQUAL 4-25
Anti Join 4-61, 4-65
APPEND 2-20, 4-68, 9-23
Architecture 1-1, 1-2, 1-3, 1-12, 1-19, 1-20, 8-14, 8-15
Ascending 4-10, 4-11, 4-15, 4-16, 4-17, 9-12, 9-14
ASH 1-14, 1-17, 1-19, 2-6, 2-8, 2-9, 3-27, 3-28, 3-30, 4-4,
4-5, 4-11, 4-24, 4-44, 4-45, 4-46, 4-47, 4-48, 4-51, 4-54, 4-58,
4-59, 4-62, 4-63, 4-64, 4-65, 4-67, 4-70, 5-11, 5-12, 5-14, 5-20,
5-21, 5-22, 5-23, 5-24, 5-25, 5-26, 5-27, 5-30, 5-32, 5-37, 5-44,
5-47, 5-49, 6-6, 6-12, 6-13, 6-16, 8-3, 9-11, 9-22, 9-29, 11-18,
11-26
Oracle Database 11g: SQL Tuning Workshop Index - 2

A
AST 1-5, 1-7, 1-16, 2-3, 2-6, 2-18, 2-19, 3-5, 3-11, 3-12,
3-16, 3-17, 3-25, 3-28, 3-29, 3-30, 3-33, 4-4, 4-5, 4-6, 4-7,
4-9, 4-10, 4-13, 4-19, 4-20, 4-21, 4-24, 4-27, 4-28, 4-29, 4-32,
4-36, 4-40, 4-45, 4-46, 4-67, 4-69, 5-4, 5-9, 5-11, 5-13, 5-19,
5-24, 5-25, 5-26, 5-27, 5-35, 5-36, 5-37, 5-38, 5-39, 5-44, 5-45,
5-46, 6-7, 6-10, 6-14, 7-16, 7-19, 7-26, 7-33, 8-3, 8-8, 8-10,
9-3, 9-7, 9-10, 9-11, 9-14, 9-15, 9-20, 9-21, 9-22, 9-23, 9-24,
9-28, 9-30, 9-31, 10-17, 10-27, 11-3, 11-5, 11-7, 11-16, 11-17, 11-18,
11-27, 11-28, 11-29, 11-30, 11-31, 11-33, 11-37
AUDIT 3-3 n s e
lice
Automatic 1-8, 1-14, 1-15, 1-16, 1-17, 1-18, 1-24, 2-6, 2-7,
a b le
2-8, 2-10, 2-11, 2-12, 2-15, 2-20, 3-30, 3-33, 4-10, 5-7, 5-10, fer
s
5-15, 5-27, 5-28, 5-29, 5-30, 5-35, 6-7, 6-10, 7-3, 7-4, 7-13, tran
o n -
a n
7-15, 7-21, 7-24, 7-25, 7-26, 7-27, 7-34, 7-36, 8-12, 8-13, 8-16,
8-19, 10-6, 10-8, 11-2, 11-3, 11-4, 11-5, 11-6, a s eฺ 11-11,
11-7, 11-8,
)h id
i r ฺes 11-30,
11-12, 11-15, 11-26, 11-27, 11-28, 11-29,
t G u 11-32, 11-33, 11-34,
11-31,
11-35, 11-36, 11-37, 11-38
g p ฺm uden
St 11-30, 11-31, 11-32,
@d11-27,s 11-28,
Automatic SQL tuning 2-6, 11-2,
11-33, 11-34, 11-35,j
ia
ab11-36, e hi
t 11-38
11-37,
(
iaOptimizer s
u11-3, 11-5, 11-6, 11-7
Automatic Tuningb t o
n
AUTOTRACE i o A5-4, 5-6, 5-15, 5-16, 5-17, 5-18, 5-19, 9-15, 10-25
n to
A
eAWR 2-6, 2-8, 2-9, 2-12, 5-4, 5-6, 5-21, 5-27, 5-28, 5-29,
Jos 5-30, 5-31, 5-32, 5-33, 5-34, 11-12, 11-14, 11-27, 11-32
awrsqrpt 2-8, 5-34
B
B*tree 3-19, 4-10, 4-11, 4-12, 4-18, 4-26, 4-27, 4-30, 4-46, 7-4,
7-8
background process 1-2, 1-3, 1-4, 1-6, 1-8, 1-15, 1-18, 1-20,
1-25, 10-12
BIND 2-14, 2-16, 2-22, 2-23, 2-24, 2-25, 3-6, 3-7, 3-11, 3-12,
3-13, 4-23, 5-14, 5-15, 6-8, 8-1, 8-2, 8-4, 8-5, 8-6, 8-7,
8-8, 8-11, 8-12, 8-13, 8-14, 8-15, 8-16, 8-17, 8-18, 8-19, 8-20,
8-21, 9-24, 10-10, 10-11, 10-12, 10-15, 10-16, 10-17, 11-12
bind peeking 5-15, 8-2, 8-13, 8-14, 8-20, 8-21

B
Bind variable 2-14, 2-16, 2-22, 2-23, 2-24, 2-25, 3-7, 5-15,
6-8, 8-1, 8-2, 8-4, 8-5, 8-6, 8-7, 8-8, 8-11, 8-12, 8-13,
8-14, 8-16, 8-18, 8-19, 8-20, 9-24, 10-17
Bind-aware 8-14, 8-15, 8-16, 8-17
Bind-sensitive 8-14, 8-16

Bitmap 1-14, 3-30, 4-4, 4-10, 4-11, 4-18, 4-20, 4-30, 4-31, 4-32,
4-33, 4-34, 4-35, 4-36, 4-39, 5-26, 6-5, 6-7, 6-8, 6-9, 6-10,
6-11, 6-13, 6-15, 6-16, 6-18, 6-19, 6-20, 6-21, 7-9, 9-12, 9-14,
9-15, 9-16
BITMAP AND 4-33, 4-34, 6-11, 6-13, 6-16, 9-16 n s e
i ce
BITMAP CONVERSION 4-31, 4-33, 4-34, 6-13, 6-16, 9-16
b l el
Bitmap create area 1-14 fe ra
n s
BITMAP INDEX 4-4, 4-10, 4-11, 4-20, 4-30, 4-31, 4-32, 4-33, 4-34,
n - tra
no
4-35, 4-39, 6-5, 6-7, 6-8, 6-9, 6-10, 6-11, 6-13, 6-15, 6-16,
a
6-18, 6-19, 6-20, 6-21, 7-9, 9-12, 9-15, 9-16 as eฺ
) h i d
Bitmap Join Index 4-11, 4-35, 6-15, 6-16, 6-18,
i r ฺes 6-19,
t G u 6-21,
6-20,
9-12
g p ฺm uden
BITMAP KEY ITERATION 4-34, d6-10, 6-13,
S t 6-16
@ i s
ia 4-34,th6-11,
BITMAP MERGE 1-14, ( j a b
3-30, e 6-13, 6-16, 9-16
i a u s
b 1-14to
Bitmap mergeAarea
BITMAP i o
nMINUS 4-34
n t o
e A
BITMAP OR 6-16, 9-16
J o s Branch 4-12, 4-20, 4-21, 4-22, 4-46, 4-68, 5-41, 5-48
Buffer cache 1-6, 1-7, 1-10, 1-11, 1-15, 1-16, 1-18, 2-15, 3-5,
3-27, 3-30, 3-32, 4-5, 4-45, 4-49, 5-18, 9-24
BUFFER SORT 4-49, 4-60, 6-13, 6-16
C
CACHE 1-6, 1-7, 1-9, 1-10, 1-11, 1-12, 1-15, 1-16, 1-18, 2-15,
3-5, 3-9, 3-14, 3-27, 3-29, 3-30, 3-31, 3-32, 4-5, 4-45, 4-49,
4-70, 4-72, 5-4, 5-18, 5-20, 5-23, 5-24, 5-25, 5-32, 7-20, 8-3,
8-4, 8-16, 8-19, 9-24, 9-25, 10-16, 10-23, 10-26, 10-28, 10-29, 11-12,
11-17, 11-21
Cardinality 2-3, 3-24, 3-26, 5-3, 5-38, 7-5, 7-12, 7-24, 7-31,
9-24, 11-7

C
Cartesian 2-3, 4-49, 4-60, 5-49, 11-11
Child 5-3, 5-20, 5-21, 5-23, 5-24, 5-25, 5-40, 5-41, 5-43, 5-44,
5-45, 5-46, 5-48, 8-12, 8-13, 8-15
Client 1-4, 1-12, 3-5, 3-8, 3-9, 3-10, 5-18, 5-19, 10-3, 10-4,
10-7, 10-8, 10-9, 10-10, 10-13, 10-14, 10-15

CLIENT_IDENTIFIER 10-8, 10-10, 10-14
CLOSE 1-5, 3-3, 3-6, 3-7, 3-12, 3-13, 3-28, 5-44, 9-5, 10-16,
10-27
Cluster 1-3, 1-22, 2-12, 2-19, 4-4, 4-6, 4-8, 4-10, 4-13,
4-44, 4-45, 4-46, 4-47, 4-56, 7-4, 7-6, 7-8, 7-9, 9-11, 10-6 n s e
i ce
Clustering factor 4-45, 4-46, 7-4, 7-6, 7-8, 7-9
b l el
COMMENT 2-24, 3-3, 9-3, 9-5, 9-6 fer a
a n s
COMMIT 1-11, 1-22, 2-13, 3-3, 3-12, 10-16, 10-18
n - tr
Composite 4-10, 4-21, 4-23, 4-28, 4-36, 4-44, 4-55, 5-37 n o
a
CONCATENATION 4-50, 4-66, 4-68, 9-17 as ฺ
)h ide
CONNECT 1-3, 1-4, 1-5, 1-6, 1-7, 1-19, 2-13,
i r ฺes2-14,t G u
2-22,
g ฺm 9-27,
2-25, 3-3, 3-5, 3-21, 4-12, 5-10,p5-15,
u d en10-6, 10-7, 10-8,
t
10-20 @d s S
ia hi
j ab 2-14,
Connection 1-4, 1-5, (2-13, e t 2-25, 10-6, 10-7, 10-8
2-22,
ia 2-20, s
u3-20, 4-10, 4-14, 4-20, 4-27, 4-35,
Constraint 2-18,
A b 2-19, t o
n
4-39,i o4-40, 7-7, 11-3
n to
A file 1-15, 1-19
econtrol
Jos CONTROL_MANAGEMENT_PACK_ACCESS 5-35, 9-25
Cost 1-14, 2-3, 2-4, 3-16, 3-18, 3-19, 3-24, 3-26, 3-27, 3-28,
3-30, 3-32, 3-33, 3-34, 4-4, 4-5, 4-7, 4-8, 4-9, 4-11, 4-14,
4-15, 4-16, 4-17, 4-18, 4-19, 4-20, 4-23, 4-25, 4-26, 4-29, 4-31,
4-34, 4-39, 4-40, 4-45, 4-46, 4-47, 4-49, 4-50, 4-51, 4-52, 4-53,
4-54, 4-55, 4-58, 4-59, 4-60, 4-64, 5-3, 5-11, 5-12, 5-14, 5-25,
5-32, 5-38, 6-3, 6-7, 6-12, 6-14, 6-17, 7-3, 7-5, 7-6, 7-8,
7-9, 7-16, 7-20, 7-31, 7-32, 9-8, 9-10, 9-12, 9-15, 9-17, 9-19,
9-24, 9-25, 9-29, 10-17, 11-7, 11-16, 11-17, 11-21, 11-22, 11-25, 11-30
COUNT STOPKEY 4-52

C
CPU 1-8, 2-3, 2-12, 3-27, 4-5, 4-8, 4-26, 4-30, 4-31, 4-47,
4-51, 4-52, 4-58, 4-60, 4-64, 5-11, 5-12, 5-14, 5-25, 5-27, 5-32,
5-35, 5-37, 5-38, 5-39, 7-3, 7-4, 7-20, 7-26, 8-13, 8-17, 9-29,
10-16, 10-24, 10-25, 10-28, 10-29, 11-4, 11-14, 11-21, 11-29, 11-30
CREATE 1-4, 1-5, 1-6, 1-8, 1-14, 1-21, 1-22, 1-24, 2-11, 2-19,
2-20, 2-24, 3-3, 3-5, 3-7, 3-8, 3-9, 3-11, 3-18, 3-21, 3-22,
3-30, 4-6, 4-10, 4-14, 4-15, 4-16, 4-17, 4-18, 4-19, 4-20, 4-23,
4-24, 4-26, 4-27, 4-28, 4-29, 4-35, 4-36, 4-38, 4-39, 4-40, 4-41,
4-44, 4-45, 4-46, 4-47, 4-51, 4-67, 5-7, 5-10, 5-15, 5-29, 5-33,
6-7, 6-14, 6-15, 6-18, 6-19, 6-20, 6-21, 7-4, 7-11, 7-13, 7-15, n s e
lice
7-16, 7-17, 7-18, 7-19, 7-21, 7-26, 7-29, 9-28, 10-7, 10-8, 10-10,
a b le
fer
10-20, 10-27, 10-29, 10-31, 11-5, 11-7, 11-8, 11-9, 11-10, 11-12, 11-14,
s
11-15, 11-16, 11-21, 11-25, 11-27, 11-28, 11-29, 11-36
- t r an
Cube 8-14, 8-15, 8-16 n on
s a3-12,
Cursor 1-14, 2-14, 2-16, 2-22, 2-23, 2-24, 2-25, 3-7,a3-8, eฺ
)h
es 5-24,
3-13, 3-29, 5-4, 5-6, 5-20, 5-21, 5-22, ฺ5-23,
id
u 5-26,
5-25,
i r t G
5-32, 5-44, 7-26, 8-2, 8-3, 8-4, p
g 8-8, en 8-11, 8-12,
ฺm8-9,ud8-10,
St 8-20, 8-21, 9-24, 10-16,
@d 8-18,s 8-19,
8-13, 8-14, 8-15, 8-16, 8-17,
10-17, 10-27, 11-12, j
ia
ab11-29 e thi
(
ia 2-25, s
u 8-2, 8-3, 8-4, 8-8, 8-9,
Cursor Sharing A b2-16, t o 3-29,
n
8-10, i o8-12, 8-13, 8-14, 8-15, 8-16, 8-17, 8-18, 8-19, 8-20, 8-21
n to
A
eCURSOR_SHARING 2-24, 3-29, 8-11, 8-12, 8-19, 8-21, 9-24, 11-29
Jos CURSOR_SHARING_EXACT 8-11, 9-24
D
Data block 1-7, 1-22, 1-23, 2-15, 4-4, 4-7, 4-12, 4-21, 4-44,
4-47, 4-49, 4-56, 5-18, 7-5, 7-6, 7-7, 7-28, 10-25, 10-29
Data dictionary cache 1-9, 3-5, 10-26
data file 1-7, 1-10, 1-11, 1-15, 1-19, 1-21, 1-23, 10-25
Database writer 1-7, 1-10, 1-15
DB_FILE_MULTIBLOCK_READ_COUNT 3-29, 4-5, 4-6, 4-20, 7-5
DBA_HIST_ACTIVE_SESS_HISTORY 5-31
DBA_HIST_BASELINE DBA_HIST_DATABASE_INSTANCE 5-31
DBA_HIST_SNAPSHOT 5-29, 5-30, 5-31
DBA_HIST_SQL_PLAN 5-4, 5-31

D
DBA_HIST_WR_CONTROL 5-30, 5-31
DBA_IND_STATISTICS 7-6, 7-8
DBA_TAB_COL_STATISTICS 3-25, 7-10
DBA_TAB_STATISTICS 3-25, 7-5, 7-34
DBA_TABLES 3-25
DBMS_APPICATION_INFO 10-4
DBMS_MONITOR 5-4, 10-4, 10-10, 10-11, 10-12, 10-13, 10-15
DBMS_SERVICE 10-4
DBMS_SESSION 10-4, 10-8, 10-10, 10-12, 10-15
DBMS_SQLTUNE 5-35, 5-37, 11-7, 11-30 n s e
i ce
DBMS_STATS 2-15, 7-3, 7-13, 7-15, 7-17, 7-18, 7-19, 7-20, 7-22,
b l el
7-23, 7-24, 7-25, 7-26, 7-28, 7-29, 7-30, 7-34, 11-6 fe ra
n s
DBMS_WORKLOAD_REPOSITORY 5-28, 5-29, 5-30, 5-33, 11-32
n - tra
DBMS_XPLAN 5-6, 5-11, 5-12, 5-14, 5-25, 5-32, 5-33, 5-34,
a no5-44,
9-29
h a s eฺ
DECLARE 3-3, 4-26, 7-16 e s ) u id
ฺ
ir 3-26, n G
t3-34,
DEFINE 1-19, 3-3, 3-6, 3-7, 3-10, 3-12, ฺ m3-22, e 4-10,
d gp Stud
4-18, 4-27, 4-29, 4-44, 4-54,
b i a@ 4-55,
t h is 6-2, 6-22, 7-5, 7-19, 7-20,
i a (ja 10-5,
8-5, 8-6, 9-24, 9-26,
u se10-7, 10-9, 10-10, 11-20, 11-31, 11-32
DELETE 1-8,A 1-19, to 3-12, 4-5, 4-18, 5-3, 5-15, 5-36,
b 3-3, 3-11,
o
ni 7-26, 7-28, 7-33, 7-34, 8-15, 9-5, 9-6, 9-28, 10-20, 10-25,
7-25,
t o
e An10-29
J os Descending 4-10, 4-11, 4-13, 4-15, 4-16, 4-17, 9-12, 9-13, 10-19
DESCRIBE 1-2, 1-25, 2-2, 2-26, 3-2, 3-3, 3-6, 3-7, 3-10,
3-35, 3-36, 4-16, 6-10, 6-11, 7-3, 7-26, 7-27, 9-21, 10-5, 11-19
DIAGNOSTIC_DEST 10-5
Dimension 2-12, 4-61, 6-3, 6-4, 6-6, 6-7, 6-8, 6-9, 6-10,
6-11, 6-12, 6-13, 6-14, 6-15, 6-16, 6-17, 6-18, 6-20, 6-21, 10-6,
11-17, 11-21
Dispatcher 1-4, 1-5
Domain 4-10, 4-11, 9-12, 9-14
DRIVING_SITE 9-22
DROP 1-8, 3-3, 4-10, 4-27, 4-37, 4-38, 4-39, 4-40, 5-10, 5-29,
5-30, 10-20, 10-27, 11-16, 11-17

D
dynamic sampling 3-25, 3-33, 3-34, 7-2, 7-17, 7-19, 7-24, 7-31,
7-32, 7-33, 7-34, 7-35, 9-24, 11-6, 11-7
Dynamic SQL 2-14, 2-16
DYNAMIC_SAMPLING 3-25, 3-32, 3-33, 7-17, 7-31, 7-32, 7-33, 9-24,
9-25
E
Equi 1-7, 1-16, 1-17, 2-2, 2-5, 2-6, 2-7, 2-8, 2-14, 2-15,
2-17, 2-18, 2-19, 2-20, 2-23, 2-25, 3-7, 3-10, 3-18, 3-19, 3-20,
3-21, 3-22, 3-23, 4-7, 4-10, 4-28, 4-30, 4-35, 4-36, 4-39, 4-40,
n s e
4-48, 4-56, 4-58, 4-59, 4-61, 4-62, 4-65, 5-3, 5-15, 5-18, 5-19,
lice
5-24, 5-34, 6-2, 6-4, 6-7, 6-19, 6-20, 6-22, 7-5, 7-6, 7-11,
a b le
7-17, 7-25, 7-27, 9-13, 9-24, 9-26, 10-20, 10-25, 10-28, 10-29, 11-4, s f er
11-8, 11-11, 11-16, 11-22, 11-27, 11-29 - t r an
Estimator 3-24, 3-25, 3-26, 3-27, 7-11 n on
s a
EXECUTE 1-10, 2-14, 2-15, 2-23, 2-25, 3-3, 3-4, h 3-6, 3-7, eฺ
a
e s ) u id
3-8, 3-9, 3-12, 3-13, 3-14, 3-19, 3-20, i r ฺ
3-24,
n t G
3-27, 4-11, 4-18,
4-50, 4-51, 4-58, 4-69, 4-70, g 5-4,
m
pฺ 5-7,t5-9, de5-10, 5-11, 5-18,
d u
S 5-41, 5-42, 5-43, 5-45,
5-20, 5-25, 5-26, 5-32, i a @ 5-38,
5-35, i s 5-40,
j ab 7-23, e th
5-46, 5-47, 5-48,( 7-22,
ia 9-11, u s 7-32, 8-4, 8-7, 8-8, 8-9, 8-10,
A
8-11, 8-14,b8-15, t o 9-22, 9-24, 9-26, 10-10, 10-16, 10-17, 10-24,
i o
ton 10-26, 10-27, 10-28, 10-29, 11-7, 11-30, 11-32, 11-36
10-25,
e An plan 2-3, 2-10, 2-11, 2-13, 2-15, 2-20, 2-21, 2-22,
Execution
J os 3-16, 3-17, 3-18, 3-19, 3-20, 3-24, 3-26, 3-28, 3-32, 3-33, 4-2,
4-14, 4-16, 4-50, 4-55, 4-56, 4-59, 4-70, 5-1, 5-2, 5-3, 5-4,
5-5, 5-6, 5-7, 5-8, 5-9, 5-10, 5-11, 5-14, 5-15, 5-16, 5-17,
5-20, 5-24, 5-25, 5-31, 5-32, 5-33, 5-38, 5-39, 5-40, 5-41, 5-42,
5-43, 5-44, 5-45, 5-46, 5-47, 5-48, 5-49, 5-50, 5-51, 5-52, 6-3,
6-6, 6-14, 7-3, 7-31, 7-35, 8-3, 8-4, 8-9, 8-10, 8-13, 8-14,
8-15, 8-18, 9-3, 9-7, 9-8, 9-10, 9-14, 9-15, 9-16, 9-20, 9-22,
9-24, 9-26, 9-29, 9-30, 9-32, 10-16, 10-17, 10-20, 10-23, 10-26, 10-27,
10-30, 11-3, 11-4, 11-6, 11-26, 11-29, 11-30
Expansion 3-18, 3-19, 4-68, 9-17

E
EXPLAIN PLAN 4-5, 4-51, 4-70, 5-3, 5-4, 5-6, 5-7, 5-8, 5-9,
5-10, 5-11, 5-12, 5-13, 5-15, 5-20, 5-40, 5-41, 5-44, 5-45, 5-50,
9-3, 9-29, 10-20, 10-27, 10-29
Expression Statistics 7-19
Extended statistics 7-4

Extent 1-21, 1-22, 1-23, 2-15, 5-48
F
FACT 2-13, 3-16, 3-29, 4-40, 4-45, 4-46, 4-61, 6-3, 6-4, 6-7,
6-8, 6-9, 6-10, 6-11, 6-12, 6-14, 6-15, 6-17, 6-18, 6-19, 7-4,
n s e
7-6, 7-8, 7-9, 7-14, 7-20, 7-29, 9-19, 9-20, 11-6, 11-7, 11-28
lice
Fact table 6-3, 6-7, 6-8, 6-9, 6-10, 6-11, 6-12, 6-14, 6-15,
a b le
6-17, 6-18, 9-19 s f er
Fast full 3-29, 3-30, 4-4, 4-9, 4-13, 4-19, 4-20, 4-24, 4-27, - t r an
4-29, 4-36, 4-69, 8-10 n on
s a
FETCH 3-3, 3-6, 3-7, 3-8, 3-10, 3-12, 3-13, 4-14, h
) a
4-15, 4-56,
i d eฺ
4-58, 4-59, 4-67, 5-17, 6-10, 7-8, 10-16,
i r ฺes10-17,
t G u 10-24, 10-25,
10-18,
10-28, 10-29 g p ฺm uden
d 4-17, S t
filter 3-25, 3-26, 4-5, 4-6, 4-14,
i a @ 4-15,
h s
i 4-18, 4-19, 4-20,
b
a 4-31, t
4-21, 4-23, 4-25,
i a (j4-29, u se 4-32, 4-45, 4-46, 4-47, 4-49, 4-50,
A
4-52, 4-55, to 4-58, 4-59, 4-60, 4-62, 4-64, 4-66, 4-67,
b4-56, 4-57,
o
ni 5-3, 5-11, 5-12, 5-25, 5-42, 5-44, 5-49, 6-6, 6-7, 6-9,
o
4-68,
t
e An6-10, 6-14, 9-22, 10-18, 11-12, 11-17, 11-21
Jos FILTER 3-25, 3-26, 4-5, 4-6, 4-14, 4-15, 4-17, 4-18, 4-19, 4-20,
4-21, 4-23, 4-25, 4-29, 4-31, 4-32, 4-45, 4-46, 4-47, 4-49, 4-50,
4-52, 4-55, 4-56, 4-57, 4-58, 4-59, 4-60, 4-62, 4-64, 4-66, 4-67,
4-68, 5-3, 5-11, 5-12, 5-25, 5-42, 5-44, 5-49, 6-6, 6-7, 6-9,
6-10, 6-14, 9-22, 10-18, 11-12, 11-17, 11-21
FIRST ROW 4-35, 4-46, 4-53, 4-58, 4-59, 4-60, 5-45, 9-10, 9-21,
9-22, 11-5
FIRST_ROWS 3-32, 3-33, 9-10, 11-7
foreground process 1-4, 5-18, 10-12
Foreign key 2-20, 4-35, 6-3, 6-4, 6-7, 6-9, 6-18
Frequency 5-28, 7-10, 7-11, 7-12, 7-13, 7-17, 7-25, 7-27

F
FULL 1-22, 2-10, 2-14, 2-15, 2-17, 3-17, 3-19, 3-28, 3-29, 3-30,
3-32, 4-3, 4-4, 4-5, 4-6, 4-9, 4-13, 4-19, 4-20, 4-24, 4-26,
4-27, 4-29, 4-34, 4-36, 4-45, 4-46, 4-47, 4-49, 4-51, 4-52, 4-53,
4-55, 4-56, 4-57, 4-58, 4-59, 4-60, 4-61, 4-62, 4-63, 4-64, 4-65,
4-67, 4-69, 4-70, 5-11, 5-12, 5-14, 5-25, 5-28, 5-30, 5-32, 5-33,
5-38, 5-39, 5-42, 5-44, 5-47, 5-49, 6-6, 6-13, 6-14, 6-16, 7-6,
7-8, 7-34, 8-5, 8-8, 8-10, 8-18, 9-7, 9-11, 9-12, 9-14, 9-19,
9-22, 9-24, 9-27, 9-29, 9-30, 10-17, 10-27, 10-28, 11-10, 11-18, 11-25,
11-27
Full Table Scan 2-14, 2-15, 3-17, 3-19, 3-29, 3-30, 3-32, 4-4, n s e
lice
4-5, 4-6, 4-9, 4-20, 4-45, 5-42, 5-44, 5-49, 6-14, 7-6, 7-8,
a b le
8-8, 8-18, 9-11, 9-12, 9-14, 9-22, 9-24, 9-29, 10-28, 11-10 fer
s
Function based 2-4, 4-10, 4-41
- t r an
no Function-based 4-6, 4-11, 4-17, 4-18, 7-19, 9-14
n
a
s eฺ
G h a
e s ) u id
ฺ
ir ent G
Gather 2-11, 4-11, 5-2, 5-3, 5-10, 5-22, 5-26, 5-51, 6-7, 7-2,
ฺ m
gp 7-15,
7-3, 7-4, 7-5, 7-6, 7-9, 7-11, 7-13,
d S tud7-17, 7-19, 7-20,
7-21, 7-22, 7-23, 7-24,i @ 7-26,
a7-25, h is 7-27, 7-28, 7-29, 7-30, 7-31,
b
a 11-28, t
7-34, 7-36, 11-5,
i a (j11-6, u se 11-29
A
Generator 3-24,b3-28 to
o
ni 1-6, 1-14, 1-22, 5-7, 5-10, 5-37, 7-25, 7-26, 9-28,
Global
t o1-3,
e An10-10, 10-11, 11-10, 11-18, 11-30
Jos GRANT 2-25, 3-3, 3-34, 5-15, 5-29
GROUP BY 2-4, 3-21, 3-25, 3-26, 4-15, 4-48, 4-51, 4-67, 4-70,
6-5, 6-6, 6-13, 6-16, 9-18, 9-22, 9-30
H
Hard parse 2-23, 3-14, 3-18, 8-7, 8-13, 8-14, 8-15, 10-23
HASH 1-14, 1-17, 3-27, 3-28, 3-30, 4-4, 4-5, 4-24, 4-44, 4-45,
4-46, 4-47, 4-48, 4-51, 4-54, 4-58, 4-59, 4-62, 4-63, 4-64, 4-65,
4-67, 4-70, 5-11, 5-12, 5-14, 5-20, 5-21, 5-22, 5-23, 5-24, 5-25,
5-26, 5-32, 5-37, 5-44, 5-47, 5-49, 6-6, 6-12, 6-13, 6-16, 8-3,
9-11, 9-22, 9-29, 11-18
Hash Cluster 4-4, 4-44, 4-45, 4-46, 4-47

H
Hash Join 1-14, 3-28, 4-5, 4-24, 4-51, 4-54, 4-58, 4-59, 4-62,
4-63, 4-64, 4-65, 5-22, 5-32, 5-47, 6-6, 6-12, 6-13, 6-16, 9-22,
9-29
Hash join area 1-14
Height-Balanced 7-14, 7-15

High-load 2-11, 10-4, 11-3, 11-4, 11-9, 11-12, 11-14, 11-26
Hint 3-24, 3-30, 4-6, 4-9, 4-17, 4-20, 4-24, 4-25, 4-26, 4-65,
4-70, 5-14, 5-20, 5-22, 5-26, 5-35, 5-47, 6-8, 6-17, 7-31, 8-11,
9-1, 9-2, 9-3, 9-4, 9-5, 9-6, 9-7, 9-8, 9-9, 9-10, 9-11,
9-12, 9-13, 9-14, 9-15, 9-16, 9-17, 9-18, 9-19, 9-20, 9-21, 9-22, n s e
i ce
9-23, 9-24, 9-25, 9-26, 9-27, 9-28, 9-29, 9-30, 9-31, 9-32, 10-26,
b l el
10-27, 11-2, 11-37 fe ra
a n s
Histogram 2-8, 2-9, 3-25, 3-26, 7-4, 7-10, 7-11, 7-12, 7-13,
o n -tr
n 8-16,
7-14, 7-15, 7-16, 7-17, 7-26, 7-29, 7-34, 8-10, 8-11, 8-14,
a
8-18
h as eฺ
Host 1-4, 2-9, 10-7 e s ) u id
ฺ
ir ent G
ฺ m
I
d gp Stud
b a@ 2-11,
Index 1-22, 1-23, 2-3, 2-4,i2-10,
t h is2-13, 2-15, 2-16, 2-17,
2-18, 2-20, 2-25, a 3-16,
(j3-11, se 3-17, 3-19, 3-21, 3-22, 3-23, 3-28,
i a u
to 4-3, 4-4, 4-5, 4-6, 4-7, 4-9, 4-10,
b3-32, 3-34,
A
3-29, 3-30,
o
o ni 4-12, 4-13, 4-14, 4-15, 4-16, 4-17, 4-18, 4-19, 4-20, 4-21,
4-11,
t
e An4-22, 4-23, 4-24, 4-25, 4-26, 4-27, 4-28, 4-29, 4-30, 4-31, 4-32,
J os 4-33, 4-34, 4-35, 4-36, 4-37, 4-38, 4-39, 4-40, 4-41, 4-42, 4-44,
4-45, 4-46, 4-47, 4-48, 4-49, 4-50, 4-51, 4-52, 4-53, 4-54, 4-55,
4-56, 4-57, 4-58, 4-59, 4-60, 4-63, 4-65, 4-67, 4-68, 4-69, 5-25,
5-42, 5-44, 5-45, 5-46, 5-49, 5-50, 6-5, 6-7, 6-8, 6-9, 6-10,
6-11, 6-13, 6-15, 6-16, 6-18, 6-19, 6-20, 6-21, 7-4, 7-6, 7-7,
7-8, 7-9, 7-11, 7-17, 7-19, 7-20, 7-24, 7-26, 7-28, 7-29, 7-30,
7-31, 7-33, 7-34, 8-3, 8-8, 8-10, 8-11, 8-18, 9-3, 9-4, 9-8,
9-9, 9-11, 9-12, 9-13, 9-14, 9-15, 9-16, 9-20, 9-21, 9-22, 9-23,
9-24, 9-25, 9-28, 9-29, 9-30, 10-14, 10-26, 10-27, 10-28, 10-29, 11-5,
11-8, 11-9, 11-10, 11-11, 11-16, 11-17, 11-18, 11-28, 11-33
Index Cluster 4-44, 4-45, 4-46, 7-8, 7-9
Index join 4-4, 4-13, 4-24, 9-13

I
Index Organized Table 4-10, 4-11
Index Scan 4-4, 4-7, 4-13, 4-16, 4-19, 4-20, 4-21, 6-16, 7-6,
9-12, 9-13, 9-14
Index-organized 1-22, 2-20, 4-4, 4-28, 4-29
INDEX_ASC 9-12, 9-14

INDEX_COMBINE 9-12, 9-14, 9-15, 9-16
INDEX_DESC 4-17, 9-12, 9-13, 9-14
INDEX_FFS 4-9, 4-20, 9-12, 9-14
INDEX_JOIN 4-24, 9-13
INDEX_SS 4-23, 9-12, 9-14 n s e
i ce
Initialization parameter 1-6, 1-18, 2-14, 2-15, 2-24, 2-25, 3-24,
b l el
3-25, 3-30, 4-20, 4-37, 4-70, 5-22, 5-35, 6-7, 7-26, 7-31, 7-32, fer a
a n s
7-35, 8-11, 8-21, 9-25, 10-5, 10-12, 10-16, 10-30
o n -tr
Inlist Iterator 4-33, 4-50
a n
s
INSERT 1-8, 3-3, 3-11, 4-39, 5-3, 5-4, 5-7, 5-8, 5-9,a5-10, eฺ
)h
es 10-19,
5-11, 5-15, 6-14, 7-33, 9-5, 9-6, 9-23, ฺ10-4,
id
u10-20, 10-25,
i r t G
10-26, 10-29, 11-36
g p ฺm uden
INTERSECT 4-48, 4-66, 4-69, d 9-5,S9-27 t
@6-11, s
j
ia
ab4-38se
Invisible 3-32, 3-34, 4-37, thi
(
ia 4-29, u
IOT 2-20, 4-11,
A b
4-28, t o 4-46
nio
J
n t o
e A pool 1-6, 1-13, 1-16, 1-18, 3-5
Java
Jo s
Join 1-14, 1-17, 2-3, 2-4, 2-19, 2-20, 2-21, 3-20, 3-21, 3-24,
3-25, 3-26, 3-28, 3-30, 4-2, 4-3, 4-4, 4-5, 4-8, 4-11, 4-13,
4-24, 4-35, 4-39, 4-45, 4-48, 4-49, 4-51, 4-54, 4-55, 4-56, 4-57,
4-58, 4-59, 4-60, 4-61, 4-62, 4-63, 4-64, 4-65, 4-69, 4-71, 5-3,
5-21, 5-22, 5-24, 5-25, 5-26, 5-32, 5-38, 5-42, 5-43, 5-45, 5-46,
5-47, 5-49, 6-3, 6-4, 6-6, 6-7, 6-8, 6-10, 6-11, 6-12, 6-13,
6-14, 6-15, 6-16, 6-18, 6-19, 6-20, 6-21, 7-3, 7-5, 7-11, 7-16,
7-17, 7-33, 9-2, 9-9, 9-10, 9-12, 9-13, 9-17, 9-20, 9-21, 9-22,
9-24, 9-27, 9-29, 9-30, 11-18
K
Key compression 4-10, 4-11

L
Large pool 1-6, 1-12, 1-16, 1-18, 11-22
LEADING 2-3, 4-11, 4-15, 4-19, 4-21, 4-23, 4-36, 6-19, 9-4,
9-20, 9-30
Leaf 4-12, 4-20, 4-21, 4-22, 4-28, 5-14, 5-40, 7-4, 7-6, 7-7,
7-8, 7-24
Library cache 1-9, 1-10, 1-11, 3-5, 3-9, 3-14, 5-4, 5-20,
5-23, 5-24, 8-3, 8-4, 10-16, 10-23, 10-26, 10-28, 10-29
Listener 1-4, 1-5, 10-6
Literal 2-4, 2-5, 2-23, 2-25, 3-29, 8-3, 8-4, 8-7, 8-9,
8-11, 8-12, 8-13, 8-18, 8-19, 9-24 n s e
i ce
Lock 1-6, 1-7, 1-9, 1-10, 1-21, 1-22, 1-23, 2-9, 2-13, 2-15,
b l el
3-12, 3-15, 3-18, 3-21, 3-22, 3-24, 3-27, 3-28, 3-29, 3-30, 3-32, fer a
a n s
4-4, 4-5, 4-6, 4-7, 4-8, 4-9, 4-11, 4-12, 4-13, 4-19, 4-20,
o n -tr
n 4-52,
4-21, 4-22, 4-28, 4-30, 4-44, 4-45, 4-46, 4-47, 4-49, 4-51,
a
as7-6, 7-7,
4-56, 4-57, 5-3, 5-12, 5-18, 5-19, 5-48, 7-4, 7-5,
)h eฺ
es 7-34,
7-8, 7-9, 7-24, 7-25, 7-28, 7-30, 7-32, ฺ7-33,
id
u 9-5,
9-4,
i r t G
9-6, 9-10, 9-17, 9-18, 9-24, 9-26,
g p ฺm
9-27, 9-28,
u d en9-29, 10-4, 10-25,
t
10-29 @d s S
Locking Statistics 7-34j abia
e thi
( us
Lookup table A bia9-24 to
6-3,
nio
M
n t o
e A
Materialized view 2-3, 2-10, 2-19, 9-17, 11-16, 11-17, 11-18
Jo s
Measures 3-24, 6-3
MERGE 1-14, 3-3, 3-11, 3-21, 3-22, 3-28, 3-30, 3-34, 4-25, 4-34,
4-48, 4-49, 4-51, 4-52, 4-54, 4-58, 4-60, 4-64, 4-65, 4-69, 5-25,
6-8, 6-10, 6-11, 6-13, 6-16, 7-33, 8-14, 8-15, 9-5, 9-6, 9-12,
9-14, 9-16, 9-17, 9-18, 9-21, 9-22, 9-24, 9-26, 9-27, 9-30, 10-13
Min/Max 4-13, 4-53
MINUS 4-32, 4-34, 4-48, 4-66, 4-69, 9-5, 9-27
MMAN 1-15, 1-18, 4-12, 5-4, 5-6, 5-7, 5-8, 5-9, 5-11, 5-12,
5-16, 5-44, 7-22, 8-5, 8-11, 8-12, 10-2, 10-8, 10-10, 10-18, 10-19,
10-23, 10-24, 10-25, 10-26, 10-27, 10-29, 10-30, 11-4, 11-8, 11-16

M
MMON 1-15, 1-21, 2-3, 2-4, 2-5, 2-6, 2-11, 2-14, 2-15, 2-25,
3-23, 3-24, 4-4, 4-10, 4-15, 4-21, 4-26, 4-44, 4-48, 4-62, 4-63,
7-20, 8-9, 10-4, 10-6, 11-26
Module 2-9, 4-37, 10-4, 10-8, 10-9, 10-10, 10-13, 10-14, 11-4, 11-21
MONITOR 1-3, 1-15, 2-6, 2-7, 2-8, 2-9, 2-10, 2-11, 5-4,
5-15, 5-19, 5-29, 5-35, 5-36, 5-37, 5-38, 5-39, 5-50, 5-52, 7-29,
8-17, 9-25, 10-4, 10-5, 10-8, 10-10, 10-11, 10-12, 10-13, 10-15, 10-29,
11-4, 11-26
Multicolumn Statistics 7-17, 7-18
n s e
N
lice
N-ary 4-3
a b le
Nested Loops 3-26, 3-32, 4-47, 4-54, 4-55, 4-56, 4-57, 4-62, 4-63, s f er
4-64, 4-65, 4-67, 5-3, 5-42, 5-43, 5-44, 5-45, 5-46, 7-5, 9-20,
- t r an
9-21 n on
s a
NO_EXPAND 9-17
) h a
i d eฺ
NO_FACT 6-17, 9-19 i r ฺes nt Gu
NO_INDEX 9-12, 9-14 g p ฺm ude
d S t
NO_MERGE 4-51, 4-52, 9-18 a @ i s
NO_MONITOR 5-35,(j9-25 abi se th
b ia to u
i o A
NO_QUERY_TRANSFORMATION 9-17
ton
NO_RESULT_CACHE 4-70, 9-25
e An
NO_USE_HASH 9-22
J os NO_USE_MERGE 9-22
NO_USE_NL 9-21
NOAPPEND 9-23
NOAUDIT 3-3
Non-equi 4-61, 4-62
NULL 3-13, 4-15, 4-19, 4-20, 4-24, 4-26, 4-27, 4-29, 4-30, 4-39,
4-41, 4-42, 4-63, 4-65, 4-68, 5-11, 5-12, 5-14, 5-25, 5-30, 5-32,
5-44, 7-4, 7-10, 7-17, 7-26, 7-27, 7-30, 7-34, 9-29
O
OPEN 1-19, 1-24, 2-14, 2-23, 3-3, 3-6, 3-8, 3-13, 4-63, 10-12,
11-15
OPT_PARAM 9-25

O
optimizer statistics 2-3, 2-11, 2-25, 7-1, 7-2, 7-3, 7-4,
7-25, 7-27, 7-28, 7-29, 7-35, 11-28
OPTIMIZER_CAPTURE_SQL_PLAN_BASELINES 3-32, 3-33, 8-19
OPTIMIZER_DYNAMIC_SAMPLING 3-25, 3-32, 3-33, 7-31, 7-32, 7-33,
9-25
OPTIMIZER_FEATURES_ENABLED 3-32, 3-35
OPTIMIZER_INDEX_CACHING 3-32, 9-25
OPTIMIZER_INDEX_COST_ADJ 3-32, 9-25
OPTIMIZER_MODE 3-32, 3-33, 9-10
OPTIMIZER_USE_INVISIBLE_INDEXES 3-32, 3-34, 4-37, 4-38 n s e
i ce
OPTIMIZER_USE_PENDING_STATISTICS 3-32, 3-34
b l el
OPTIMIZER_USE_SQL_PLAN_BASELINES 3-32, 3-33 fer a
ORDER BY 4-15, 4-16, 4-17, 4-18, 4-19, 4-46, 4-48, 7-13, 7-15, tr a n s
o n -
9-30, 10-19
a n
s eฺ
ORDERED 3-12, 4-19, 4-20, 4-21, 4-23, 4-46, 4-48,a5-41, 5-47,
)h id
7-6, 9-20, 9-21, 9-23, 9-24, 11-21, 11-25
i r ฺes nt G u
ORDERED_PREDICATES 9-23, 9-24
g p ฺm ude
t
OS block 1-21 @d s S
Outer Join 4-61, 4-63(jab
ia
e thi
ia u s
P A b t o
i o
ton 1-3, 4-6, 7-26, 7-29, 7-30
parallelism
e An
PARALLELIZE 3-6, 3-7, 3-11, 4-40, 7-26
J os Parent 2-21, 3-20, 4-65, 5-3, 5-10, 5-23, 5-24, 5-40, 5-41, 5-43,
5-45, 5-46, 5-48, 9-5, 11-8
PARSE 1-10, 2-14, 2-23, 2-25, 3-5, 3-6, 3-7, 3-9, 3-13, 3-14,
3-15, 3-18, 5-23, 7-20, 8-4, 8-7, 8-8, 8-13, 8-14, 8-15, 10-16,
10-17, 10-18, 10-23, 10-24, 10-25, 10-26, 10-27, 10-28, 10-29
Partition 1-22, 2-3, 2-10, 2-19, 3-11, 4-3, 4-4, 5-3, 5-12,
6-8, 6-13, 7-26, 7-28, 7-30, 9-12, 9-13, 9-14, 11-16, 11-17, 11-18
Peeking 5-15, 8-2, 8-7, 8-8, 8-13, 8-14, 8-20, 8-21
PGA 1-6, 1-14, 1-15, 1-17, 1-18, 3-5, 3-29, 3-30, 4-49, 5-26,
6-10
PGA_AGGREGATE_TARGET 1-17, 3-29, 3-30, 5-26, 6-10

P
PLAN_TABLE 5-4, 5-6, 5-7, 5-10, 5-11, 5-12, 5-13, 5-14, 5-15,
5-20, 5-21, 5-25, 5-32, 5-33, 5-40, 10-20
PLUSTRACE 5-15
Predicate 2-3, 2-25, 3-18, 3-22, 3-25, 3-26, 3-27, 3-34, 4-5,
4-11, 4-14, 4-15, 4-16, 4-17, 4-18, 4-19, 4-20, 4-23, 4-24, 4-25,
4-26, 4-29, 4-30, 4-31, 4-32, 4-33, 4-34, 4-41, 4-49, 4-50, 4-51,
4-52, 4-53, 4-54, 4-55, 4-58, 4-59, 4-60, 4-61, 4-64, 4-67, 4-68,
5-11, 5-12, 5-14, 5-25, 5-42, 5-49, 6-6, 6-10, 6-14, 6-16, 7-3,
7-6, 7-10, 7-14, 7-16, 7-17, 7-18, 7-19, 7-24, 7-31, 7-32, 8-14,
8-15, 8-16, 9-15, 9-21, 9-23, 9-24, 9-26, 11-7, 11-11, 11-18 n s e
i ce
Preference 7-2, 7-25, 7-26, 7-35
b l el
PREPARE 3-3 fer a
Primary key 2-20, 3-15, 3-20, 4-11, 4-14, 4-27, 4-28, 4-29, 4-35, tra n s
o n -
4-40, 5-37, 5-46, 6-3, 6-9, 6-18, 7-7, 8-9
a n
h Private SQL area 3-4, 3-8, 3-14 a s eฺ
e s ) u id PUSH_PRED 9-24
ฺ
ir ent G
ฺ m PUSH_SUBQ 9-24
d gp Stud
Q
b i a@ this
(ja3-22,us3-24,
Query block 3-18, 3-21,
i a
e 3-28, 4-51, 5-12, 9-4, 9-17,
A
9-18, 9-26, to 9-29
b9-27, 9-28,
ni o
R n t o
e A
Jos
Range 3-25, 3-30, 3-32, 4-4, 4-5, 4-6, 4-8, 4-10, 4-13, 4-15,
4-16, 4-17, 4-18, 4-25, 4-26, 4-28, 4-31, 4-34, 4-36, 4-39, 4-49,
4-50, 4-53, 4-56, 4-57, 4-63, 4-65, 4-67, 4-68, 5-29, 5-30, 5-42,
5-44, 5-45, 6-13, 6-16, 7-8, 7-9, 7-10, 7-12, 7-14, 7-26, 8-8,
8-9, 8-10, 8-16, 8-18, 9-8, 9-12, 9-13, 9-14, 9-16, 9-29, 10-26,
10-29, 11-14, 11-18
Redo log buffer 1-6, 1-8, 1-10, 1-11
redo log file 1-8, 1-11, 1-15, 1-19, 2-15
RENAME 3-3, 5-10
Result cache 1-9, 1-15, 3-5, 3-30, 3-31, 4-70, 4-72, 9-25
Result Set 2-5, 4-54, 4-70, 5-44, 5-47, 6-7, 6-9, 6-11, 6-12,
6-14
RESULT_CACHE 3-29, 3-30, 3-31, 4-70, 9-25

R
RESULT_CACHE_MAX_RESULT 3-29, 3-31
RESULT_CACHE_MAX_SIZE 3-29, 3-31
RESULT_CACHE_MODE 3-29, 3-30, 4-70, 9-25
RESULT_CACHE_REMOTE_EXPIRATION 3-29, 3-31
Reverse key 4-10, 4-11

REVOKE 3-3
REWRITE 2-11, 2-27, 3-18, 6-9, 9-17, 11-11, 11-17
ROLLBACK 1-10, 2-13, 3-3, 3-12, 10-16, 10-18
Root 4-12, 4-20, 4-21, 5-40, 7-6, 10-5
Row cache 1-9 n s e
i ce
ROWID 4-4, 4-6, 4-7, 4-12, 4-13, 4-14, 4-15, 4-16, 4-17, 4-18,
b l el
fe
4-19, 4-23, 4-24, 4-25, 4-28, 4-30, 4-31, 4-33, 4-34, 4-36, 4-46, ra
n s
n - tra
4-47, 4-49, 4-50, 4-55, 4-56, 4-57, 4-59, 4-63, 4-68, 5-25, 5-42,
o
5-44, 5-45, 5-46, 6-11, 6-13, 6-15, 6-16, 7-5, 7-8, 9-8,n9-11,
a
s eฺ
9-16, 10-26, 10-29
h a
ROWID Scan 4-4, 4-7 e s ) u id
ฺ
ir 11-16 n tG
ฺ
Rule 2-12, 2-24, 5-41, 5-42, 5-44, 9-6, m
11-11, e
d gp Stud
S
b i a@ this
Sample Table Scan 4-4,
i a (ja 4-8,u4-9
se
SAVEPOINT A b 3-12 to
3-3,
o
ni 1-17, 2-12, 2-13
o
Scalability
t
e An 1-21, 1-22, 2-13, 2-15, 2-18, 2-24, 2-25, 3-3, 3-16, 3-17,
Schema
J os 3-24, 5-10, 5-15, 5-20, 6-2, 6-3, 6-4, 6-5, 6-6, 6-22, 7-25,
7-26, 7-27, 7-28, 7-29, 7-30, 7-34, 9-11, 9-21, 10-19, 10-20, 11-12,
11-16, 11-21, 11-38
Segment 1-15, 1-21, 1-22, 1-23, 2-8, 2-15, 3-29, 4-12, 4-20,
4-28, 4-29, 5-27, 5-48, 10-25

S
SELECT 1-4, 2-3, 2-4, 2-5, 2-11, 2-19, 2-20, 3-3, 3-10, 3-11,
3-12, 3-13, 3-16, 3-17, 3-19, 3-20, 3-21, 3-22, 3-23, 3-24, 3-25,
3-26, 3-28, 3-32, 4-3, 4-5, 4-6, 4-7, 4-8, 4-11, 4-14, 4-15,
4-16, 4-17, 4-18, 4-19, 4-20, 4-21, 4-23, 4-24, 4-25, 4-26, 4-27,
4-29, 4-31, 4-32, 4-33, 4-36, 4-38, 4-39, 4-41, 4-47, 4-48, 4-49,
4-50, 4-51, 4-52, 4-53, 4-54, 4-55, 4-56, 4-57, 4-58, 4-59, 4-60,
4-62, 4-63, 4-64, 4-65, 4-67, 4-68, 4-70, 5-3, 5-6, 5-9, 5-10,
5-11, 5-12, 5-14, 5-15, 5-18, 5-25, 5-32, 5-33, 5-34, 5-37, 5-38,
5-42, 5-44, 5-45, 5-47, 5-48, 5-49, 6-5, 6-6, 6-9, 6-12, 6-14,
6-16, 6-18, 6-19, 6-20, 6-21, 7-3, 7-6, 7-8, 7-10, 7-11, 7-13, n s e
i ce
7-14, 7-15, 7-16, 7-17, 7-18, 7-19, 7-24, 7-31, 7-32, 7-34, 8-3,
b l el
8-4, 8-5, 8-7, 8-8, 8-9, 8-10, 8-12, 8-14, 8-15, 8-16, 8-18, fe r a
a
9-3, 9-5, 9-6, 9-8, 9-10, 9-11, 9-12, 9-13, 9-14, 9-15, 9-16, tr n s
o n -
a n
9-18, 9-20, 9-21, 9-22, 9-23, 9-24, 9-27, 9-28, 9-29, 9-30, 10-7,
10-12, 10-13, 10-15, 10-17, 10-20, 10-25, 10-28, s
a 10-29,eฺ11-7, 11-8, 11-12,
)h id
11-13, 11-14, 11-15, 11-19, 11-20, 11-24,
i r ฺes 11-25,
t G u 11-34, 11-36
11-33,
n 6-14,
Selectivity 2-3, 2-11, 3-24, 3-25, 3-26,
g p ฺm
4-3, 4-6,
u d e4-36,
d 7-17, t
7-3, 7-8, 7-10, 7-11, 7-14,
ia @7-16,
hi
t 11-7 s S 7-19, 7-24, 7-31, 7-32,
j ab 8-18,
8-8, 8-14, 8-15, (8-16, s e9-24,
Semi join 4-61, b ia to u
4-64
n io A1-4, 1-5, 1-6, 1-7, 1-8, 1-9, 1-10, 1-11,
serveroprocess
t
n1-12,
e A 3-11, 3-30, 10-3
Jos Service 1-4, 1-20, 2-8, 2-9, 5-31, 10-3, 10-4, 10-6, 10-7,
10-8, 10-9, 10-10, 10-13, 10-14, 10-31
SERVICE_NAME 10-7, 10-8, 10-14
SERVICE_NAMES 10-8
Session 1-4, 1-5, 1-6, 1-12, 1-13, 1-14, 1-17, 2-8, 2-9,
3-3, 3-30, 3-33, 4-37, 5-20, 5-21, 5-25, 5-27, 5-30, 5-31, 5-35,
5-36, 5-37, 5-38, 5-39, 5-44, 7-33, 8-5, 8-11, 8-12, 9-6, 9-10,
9-27, 10-2, 10-3, 10-4, 10-6, 10-8, 10-10, 10-11, 10-12, 10-13, 10-14,
10-15, 10-17, 10-18, 10-20, 10-30, 11-19, 11-20
SESSION_ID 10-8, 10-11, 10-14
SET ROLE 3-3
SET TRANSACTION 3-3

S
SGA 1-3, 1-4, 1-6, 1-7, 1-8, 1-9, 1-10, 1-11, 1-12, 1-15,
1-16, 1-18, 2-6, 2-8, 3-5, 3-30, 3-31, 4-49, 11-21
SGA_TARGET 1-16
Shared pool 1-6, 1-9, 1-10, 1-11, 1-12, 1-15, 1-16, 1-18, 2-14,
2-22, 2-23, 3-5, 4-70, 5-24

Shared server 1-4, 1-12, 1-14, 6-10, 10-3
Shared SQL area 1-10, 3-4, 3-5, 3-9, 3-14, 3-15, 5-23
Single-table hash cluster 4-46
Skip 4-4, 4-13, 4-21, 4-22, 4-23, 4-36, 4-58, 9-14, 9-17
Small table 4-6, 4-39, 5-49 n s e
i ce
Snapshot 2-6, 5-27, 5-28, 5-29, 5-30, 5-31, 5-33, 5-34, 11-12,
b l el
11-14, 11-32 fer a
a n s
Snowflake 6-4, 6-21
o n -tr
Soft parse 2-23, 3-14, 8-14, 8-15, 10-23
a n
s eฺ
SORT 1-14, 1-17, 2-5, 2-20, 3-26, 3-27, 3-28, 3-30,a4-11, 4-15,
)h id
4-17, 4-19, 4-20, 4-24, 4-27, 4-31, 4-46,
i r ฺes4-48,t G u 4-52, 4-53,
4-49,
g p d en5-19, 5-22, 5-24,

ฺm5-3,u5-18,
4-54, 4-58, 4-60, 4-65, 4-67, 4-69,
St 9-10, 9-22, 10-18, 10-19,
@d 6-16,s 9-3,
5-25, 5-26, 5-38, 5-47, 6-13,
j
10-21, 10-22, 10-28,
ia
ab10-29e t hi
(
ia to u s
Sort area 1-14,
A b4-49
n
Sort-mergei oJoin 4-48, 4-54, 4-58, 9-22
n to
A hash cluster 4-46
eSorted
Jos SPA 1-4, 1-5, 1-8, 1-21, 1-22, 1-23, 1-24, 2-6, 2-8, 2-9,
2-14, 2-15, 2-21, 2-24, 2-25, 3-5, 3-9, 4-5, 4-10, 4-11, 4-35,
4-39, 4-40, 4-46, 4-65, 5-4, 5-5, 5-12, 5-31, 5-34, 5-36, 5-39,
5-48, 6-4, 7-5, 7-14, 7-21, 7-28, 9-3, 9-5, 9-6, 9-23, 10-8,
10-14, 10-16, 10-26, 11-4, 11-8, 11-22, 11-25
SQL Access Advisor 2-10, 5-50, 11-2, 11-10, 11-16, 11-17, 11-18, 11-19,
11-20, 11-21, 11-22, 11-23, 11-24, 11-25, 11-37, 11-38
SQL Management Base 3-33, 5-4, 5-5
SQL monitor 2-10, 5-35, 5-36, 5-37, 5-38, 5-39, 5-50, 5-52, 9-25
SQL profile 2-6, 5-5, 11-5, 11-7, 11-8, 11-9, 11-15, 11-26, 11-27,
11-28, 11-29, 11-30, 11-32, 11-33, 11-36

S
SQL tuning advisor 2-6, 2-10, 2-14, 5-50, 11-2, 11-3, 11-4, 11-5,
11-8, 11-9, 11-12, 11-13, 11-14, 11-15, 11-26, 11-27, 11-28, 11-34, 11-36,
11-37, 11-38
SQL tuning loop 11-9, 11-26
SQL tuning set 2-10, 5-4, 5-6, 11-12, 11-14, 11-17, 11-21, 11-32
Stale 2-3, 3-31, 7-3, 7-5, 7-6, 7-25, 7-26, 7-27, 11-5, 11-6,
11-28, 11-29
Star Join 4-61
star query 6-2, 6-3, 6-5, 6-7, 6-12, 6-14, 6-16, 6-17, 6-22
star schema 6-2, 6-3, 6-4, 6-6, 6-22 n s e
i ce
Star Transformation 3-18, 4-35, 6-1, 6-2, 6-5, 6-6, 6-7,
b l el
6-8, 6-9, 6-12, 6-13, 6-14, 6-15, 6-16, 6-17, 6-22, 6-23, 9-17, fer a
a n s
9-19
o n -tr
n 9-25
STAR_TRANSFORMATION 3-29, 3-30, 6-7, 6-17, 9-4, 9-19,
a
STAR_TRANSFORMATION_ENABLED 3-29, 3-30,a6-7, s 9-25eฺ
Statistics 1-16, 1-18, 2-3, 2-6, 2-7, 2-8, 2-9,ฺe
)h
s 2-15,
2-11,
id
u
i r t G
2-25, 3-16, 3-17, 3-18, 3-24, 3-25,
g p en 3-33, 3-34, 4-11,
ฺm3-26,ud3-32,
@d5-16,s5-17,
4-41, 4-42, 5-4, 5-13, 5-15, St 5-18, 5-19, 5-20, 5-22,
5-23, 5-24, 5-26, j
ia
ab 5-28,
5-27, e hi
t 5-31, 5-35, 5-37, 5-44, 5-48,
5-29,
( s
u 7-6, 7-7, 7-8, 7-9, 7-10,
6-7, 7-1,A bia7-3, 7-4,
7-2, to 7-5,
n
7-11,
o io7-12, 7-13, 7-15, 7-17, 7-18, 7-19, 7-20, 7-21, 7-22, 7-23,
t
e An7-24, 7-25, 7-26, 7-27, 7-28, 7-29, 7-30, 7-31, 7-34, 7-35, 7-36,
Jos 8-10, 8-11, 8-14, 8-15, 8-17, 8-18, 9-3, 9-10, 9-24, 10-2, 10-10,
10-16, 10-17, 10-18, 10-20, 10-23, 10-24, 10-25, 10-26, 10-28, 10-30, 11-5,
11-6, 11-7, 11-8, 11-9, 11-12, 11-14, 11-15, 11-24, 11-26, 11-28, 11-29,
11-32, 11-33
STATISTICS_LEVEL 1-16, 1-18, 5-22, 5-26, 5-35, 5-44, 7-27, 10-16,
11-32
STATS$SQL_PLAN 5-4
Streams pool 1-6, 1-13, 1-16, 1-18, 3-5
Suboptimal 2-3, 2-21, 7-11, 7-19, 8-8, 11-6
Subquery 2-4, 3-11, 3-18, 3-20, 4-61, 4-64, 4-65, 5-42, 6-8,
7-33, 9-5, 9-17, 9-18, 9-24, 9-27
SYSAUX 1-24, 5-5

S
SYSTEM 1-3, 1-4, 1-6, 1-8, 1-11, 1-15, 1-16, 1-17, 1-18, 1-23,
1-24, 2-8, 2-9, 2-11, 2-12, 2-13, 2-14, 2-15, 2-16, 2-22, 2-23,
2-25, 3-3, 3-18, 3-26, 3-30, 3-31, 4-3, 4-8, 4-10, 4-14, 4-15,
4-16, 4-21, 4-22, 4-23, 4-25, 4-29, 4-30, 4-35, 4-37, 4-45, 4-47,
4-49, 4-52, 4-57, 4-59, 4-63, 5-10, 5-17, 5-18, 5-27, 5-28, 5-31,
5-44, 5-46, 5-47, 6-9, 6-14, 7-2, 7-3, 7-4, 7-11, 7-13, 7-15,
7-20, 7-21, 7-22, 7-23, 7-24, 7-26, 7-27, 7-28, 7-29, 7-33, 7-35,
7-36, 8-8, 8-11, 8-12, 8-14, 8-15, 8-19, 9-3, 9-22, 9-24, 9-29,
10-6, 10-8, 10-9, 10-16, 10-20, 11-3, 11-6, 11-8, 11-9, 11-14, 11-17,
11-26, 11-34, 11-36 n s e
lice
System statistics 2-11, 7-2, 7-4, 7-20, 7-21, 7-22, 7-23, 7-27,
a b le
7-28, 7-35, 7-36, 9-24 fer s
T - t r an
n
no
Table 1-9, 1-21, 1-22, 1-23, 1-24, 2-3, 2-5, 2-14, 2-15, 2-16, a
2-17, 2-18, 2-19, 2-20, 2-21, 2-25, 3-10, 3-11, h a s eฺ 3-17,
3-12, 3-16,
e s ) u id
3-18, 3-19, 3-20, 3-21, 3-22, 3-25, 3-26,
i r ฺ 3-28,
n t G 3-30, 3-31,
3-29,
m
ฺ 4-7,ud4-8, e 4-9, 4-10,
3-32, 3-33, 3-35, 4-3, 4-4, 4-5,
d gp4-6, t
S 4-18, 4-19, 4-20, 4-21,
4-11, 4-12, 4-13, 4-14,i a @ 4-16,
4-15, i s 4-17,
j ab 4-27, e th
(
4-23, 4-24, 4-25, 4-26,
ia 4-37, u s 4-28, 4-29, 4-30, 4-31, 4-32, 4-33,
A b
4-34, 4-35, 4-36, t o 4-39, 4-40, 4-44, 4-45, 4-46, 4-47, 4-49,
i o
n ton 4-51, 4-52, 4-54, 4-55, 4-56, 4-57, 4-58, 4-59, 4-60, 4-61,
4-50,
e A 4-62, 4-63, 4-64, 4-65, 4-67, 4-68, 4-70, 5-3, 5-4, 5-5, 5-6,
J o s
5-7, 5-8, 5-9, 5-10, 5-11, 5-12, 5-13, 5-14, 5-15, 5-18, 5-20,
5-21, 5-25, 5-31, 5-32, 5-33, 5-34, 5-38, 5-39, 5-40, 5-42, 5-44,
5-45, 5-46, 5-47, 5-48, 5-49, 6-3, 6-4, 6-5, 6-6, 6-7, 6-8,
6-9, 6-10, 6-11, 6-12, 6-13, 6-14, 6-15, 6-16, 6-17, 6-18, 6-20,
6-21, 7-4, 7-5, 7-6, 7-7, 7-8, 7-9, 7-13, 7-15, 7-17, 7-18,
7-19, 7-20, 7-21, 7-22, 7-24, 7-25, 7-26, 7-27, 7-28, 7-29, 7-30,
7-31, 7-32, 7-33, 7-34, 8-4, 8-8, 8-9, 8-10, 8-11, 8-18, 9-3,
9-4, 9-5, 9-6, 9-8, 9-9, 9-10, 9-11, 9-12, 9-13, 9-14, 9-16,
9-19, 9-20, 9-21, 9-22, 9-23, 9-24, 9-26, 9-27, 9-28, 9-29, 9-30,
10-5, 10-17, 10-18, 10-19, 10-20, 10-21, 10-22, 10-25, 10-26, 10-28, 10-29,
11-5, 11-6, 11-8, 11-9, 11-10, 11-17, 11-18, 11-21, 11-25, 11-36

T
Tablespace 1-21, 1-22, 1-23, 1-24, 2-15, 2-25, 4-39, 4-40, 5-5,
5-31, 5-48, 7-21, 11-25
Temporary table 1-21, 1-22, 4-49, 5-7, 5-10, 6-14, 7-31
tkprof 2-8, 5-10, 5-48, 7-35, 10-2, 10-4, 10-13, 10-17, 10-18, 10-19,
10-20, 10-21, 10-22, 10-23, 10-24, 10-25, 10-26, 10-27, 10-28, 10-29, 10-30,
10-31
trace file 1-19, 1-20, 2-6, 2-9, 3-28, 3-37, 5-4, 5-5, 5-14,
5-40, 5-47, 5-48, 7-35, 10-2, 10-3, 10-5, 10-8, 10-10, 10-12, 10-13,
10-14, 10-15, 10-16, 10-17, 10-18, 10-19, 10-20, 10-27, 10-30
Transformer 3-18, 3-19, 3-20, 3-21, 3-22, 3-23 n s e
lice
Transitivity 3-18, 3-23
a b le
trcsess 2-8, 10-2, 10-4, 10-13, 10-14, 10-15, 10-18, 10-30, 10-31
sfer
TRUNCATE 3-3
- t r an
no Tuning mode 11-3
n
a
s eฺ
U h a
e s ) u id
UGA 4-49 ฺ
ir ent G
ฺ m
d gp 4-50,
UNION 2-3, 2-4, 2-5, 3-19, 3-22, 4-48,
S
d 4-68, 4-69,
tu4-66,
5-48, 9-5, 9-17, 9-27,ia @ 11-11
9-28,
h is
b
ja 4-50, t
UNION ALL 2-3, 2-5,(3-19,
i a u se4-66, 4-68, 4-69, 5-48, 9-17,
A
9-27, 11-11b to
o
i 2-5, 2-20, 3-8, 3-20, 4-4, 4-10, 4-11, 4-13, 4-14,
n1-11,
UNIQUE
t o
e An4-15, 4-27, 4-29, 4-39, 4-40, 4-48, 4-50, 4-55, 4-56, 4-60, 4-69,
Jos 5-21, 5-24, 5-37, 5-42, 5-44, 5-45, 5-46, 7-7, 7-9, 7-16, 8-11,
9-8, 10-4
Unique 1-11, 2-5, 2-20, 3-8, 3-20, 4-4, 4-10, 4-11, 4-13, 4-14,
4-15, 4-27, 4-29, 4-39, 4-40, 4-48, 4-50, 4-55, 4-56, 4-60, 4-69,
5-21, 5-24, 5-37, 5-42, 5-44, 5-45, 5-46, 7-7, 7-9, 7-16, 8-11,
9-8, 10-4
Unique key 2-20, 3-20, 4-40
UNNEST 3-18, 3-20, 3-34, 4-64, 9-4, 9-17
UPDATE 1-8, 3-3, 3-11, 3-12, 4-7, 4-37, 5-3, 5-15, 5-23, 7-3,
7-20, 7-21, 7-22, 7-33, 9-5, 9-6, 9-8, 9-28, 10-4, 10-25, 10-29
USE_CONCAT 9-17
USE_HASH 5-47, 9-22

U
USE_MERGE 9-21, 9-22, 9-30
USE_NL 5-45, 9-4, 9-20, 9-21, 9-22, 9-30
USE_NL_WITH_INDEX 9-21
user process 1-3, 1-4, 1-5, 1-11, 1-15, 3-9, 3-14
V
V$ACTIVE_SESSION_HISTORY 5-31, 5-35, 5-36, 5-38
V$DIAG_INFO 10-5
V$SESSION 5-21, 5-25, 5-35, 5-36, 5-39, 10-8, 10-11, 10-14
V$SESSION_LONGOPS 5-35, 5-36, 5-39
n s e
V$SQL 5-4, 5-13, 5-14, 5-20, 5-21, 5-22, 5-23, 5-24, 5-25, 5-26,
lice
5-32, 5-35, 5-36, 5-37, 5-38, 5-40, 5-44, 5-48, 8-3, 8-14, 8-16,
a b le
8-17, 11-17 s f er
V$SQL_CS_HISTOGRAM 8-16 - t r an
V$SQL_CS_SELECTIVITY 8-16 n on
s a
V$SQL_CS_STATISTICS 8-17
) h a
i d eฺ
V$SQL_MONITOR 5-4, 5-35, 5-36, 5-37 irฺe
s Gu
V$SQL_PLAN 5-4, 5-13, 5-14, 5-20, p ฺ m 5-22,
5-21, d e nt 5-24, 5-25,
5-23,
d g Stu
5-26, 5-35, 5-36, 5-38,
i @ 5-44,
a5-40, h is 5-48
b t
V$SQL_PLAN_MONITOR
i a (ja 5-4, u se5-35, 5-36, 5-38
A b
V$SQL_PLAN_STATISTICS to 5-13, 5-20, 5-22, 5-23, 5-24, 5-48
i o
ton
V$SQL_PLAN_STATISTICS_ALL 5-13, 5-20, 5-22, 5-23, 5-24
e An
V$SQL_WORKAREA 5-20, 5-22, 5-23, 5-24
J os V$SQLAREA 5-21, 5-23, 5-24, 5-25
View 1-2, 1-3, 1-6, 1-20, 1-22, 1-26, 2-3, 2-6, 2-8, 2-9,
2-10, 2-11, 2-16, 2-19, 2-21, 2-25, 2-27, 3-6, 3-14, 3-15, 3-18,
3-21, 3-22, 3-25, 3-26, 3-32, 3-34, 3-37, 4-8, 4-10, 4-24, 4-35,
4-37, 4-38, 4-43, 4-50, 4-51, 4-52, 4-54, 4-72, 5-4, 5-6, 5-13,
5-15, 5-20, 5-21, 5-22, 5-23, 5-24, 5-27, 5-29, 5-30, 5-31, 5-32,
5-35, 5-36, 5-38, 5-39, 5-49, 5-52, 6-6, 6-8, 6-15, 6-18, 6-23,
7-4, 7-9, 7-11, 7-13, 7-15, 7-17, 7-18, 7-19, 7-25, 7-26, 7-28,
7-31, 7-33, 7-34, 7-35, 7-36, 8-13, 8-14, 8-16, 8-17, 8-21, 9-3,
9-4, 9-17, 9-18, 9-24, 9-25, 9-26, 9-27, 9-28, 9-29, 9-31, 9-32,
10-6, 10-9, 10-12, 10-14, 10-18, 10-30, 10-31, 11-5, 11-8, 11-15, 11-16,
11-17, 11-18, 11-19, 11-20, 11-23, 11-24, 11-25, 11-28, 11-33, 11-34, 11-38

V
View merging 3-18, 3-21, 3-34, 9-17, 9-18
W
Wait 2-8, 2-9, 2-12, 2-15, 5-33, 5-35, 5-37, 5-38, 10-10, 10-11,
10-12, 10-15, 10-16, 10-17, 10-19, 11-36
WHENEVER 3-3, 4-62, 7-3, 7-25
n s e
i ce
b l el
fer a
a n s
n - tr
o
s an
) h a deฺ
r ฺ e s Gui
p ฺ mi dent
d g Stu
b i a@ this
i a (ja use
Ab to
o n io
n t
e A
J o s

D52163GC10 sg2

Uploaded by

Copyright:

Available Formats

D52163GC10 sg2

Uploaded by

Document Information

Original Title

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

D52163GC10 sg2

Uploaded by

Copyright:

Available Formats

Unauthorized reproduction or distribution prohibitedฺ Copyright© 2010, Oracle and/or its affiliatesฺ

Jean-François Verrier Disclaimer

This document contains proprietary information and is protected by copyright and

1 Exploring the Oracle Database Architecture

J o s SYSTEM and SYSAUX Tablespaces 1-24

2 Introduction to SQL Tuning

Table Design 2-19

J o s SQL Statement Parsing: Overview 3-14

Buffer Sort Operator 4-49

5 Interpreting Execution Plans

Using the V$SQL_PLAN View 5-20

Bitmap Join Indexes: Join Model 2 6-19

Bind Variables in SQL*Plus 8-5

t o ni Hint Recommendations 9-7

Location for Diagnostic Traces 10-5

SQL Access Advisor Session: Initial Options 11-19

Practice 2-1: Avoiding Common Mistakes................................................................... 13

Oracle Database 11g: SQL Tuning Workshop A - 2

Oracle Database 11g: SQL Tuning Workshop A - 3

2) An instance consists of _____________________and _____________________

3) A session is a connection between the _______________ process and either the

8) Name some files associated with an Oracle database.

Oracle Database 11g: SQL Tuning Workshop A - 4

9) Some of the logical storage structures of an Oracle database are: e

12) The logical tablespace structure is associated with the physical

13) State whether the following statements are true or false.

Oracle Database 11g: SQL Tuning Workshop A - 5

machine. This includes both background processes and foreground processes.

Oracle Database 11g: SQL Tuning Workshop A - 6

SQL*Plus: Release 11.1.0.6.0 - Production on Thu Mar 27 17:58:56

Copyright (c) 1982, 2007, Oracle. All rights reserved.

Oracle Database 11g: SQL Tuning Workshop A - 7

Oracle Database 11g: SQL Tuning Workshop A - 8

set echo on (ja se

select group#,member from v$logfile;

col tablespace_name format a20

select tablespace_name, file_name from dba_data_files;

select tablespace_name, file_name from dba_temp_files;

Oracle Database 11g: SQL Tuning Workshop A - 9

SQL*Plus: Release 11.1.0.6.0 - Production on Thu Mar 27 18:15:02

Copyright (c) 1982, 2007, Oracle. All rights reserved.

COMPONENT CURRENT_SIZE MIN_SIZE MAX_SIZE

Oracle Database 11g: SQL Tuning Workshop A - 10

COMPONENT CURRENT_SIZE MIN_SIZE MAX_SIZE

ASM Buffer Cache 0 0 0

select * from v$sgainfo;

col component format a30

select component,current_size,min_size,max_size from

Oracle Database 11g: SQL Tuning Workshop A - 11

Oracle Database 11g: SQL Tuning Workshop A - 12

SQL*Plus: Release 11.1.0.6.0 - Production on Wed Mar 26 20:21:50

sqlplus / as sysdba <<EOF

alter user sh identified by sh account unlock;

grant dba to sh;

Oracle Database 11g: SQL Tuning Workshop A - 13

SQL*Plus: Release 11.1.0.6.0 - Production on Wed Mar 26 20:37:49

alter system flush shared_pool;

Oracle Database 11g: SQL Tuning Workshop A - 14

Oracle Database 11g: SQL Tuning Workshop A - 15

2) An instance consists of _and _