Proposed Outline For Physical Database Design by Wells Fargo Day 1

Download as doc, pdf, or txt
Download as doc, pdf, or txt
You are on page 1of 4

Proposed Outline for Physical Database Design by Wells Fargo

Day 1
1. Building the Database Environment
2. Data Distribution, Hashing, and Index Aess
!. "hoosing a #rimary Index
Day 2
1. $eondary Index %sage
2. $eondary Indexes and Further "onsiderations
!. &o' and (able $trutures
Day 3
1. #artitioned #rimary Indexes
2. Database Integrity
!. "olleting $tatistis
Day 4
1. )oin #roessing
2. Database Design "hoies
!. (able Design *uidelines
Teradata Physical Database Design
This course is a Customized Course for Wells Fargo to teach how to defines the processes and
procedures to follow when designing a database for implementation on a Teradata system. It
explains how to transform a logical data model into an extended data model. The focus of this
course is on the C!"T! T"#$! statement% including all aspects of table design. It co&ers Teradata
data distribution% access% and use of deri&ed data. 'imilarities between (oin and aggregation
processing% and the implementation of eferential Integrity are also discussed. "lso co&ered in the
course are the steps to optimize space management. ')$ *ueries
DAY 1 & 2 :
+odule ,- Physical Database Design .&er&iew
Describe the stages of database design.
$ist and describe the input re*uirements for database design.
$ist and describe the outputs and ob(ecti&es for database design.
Describe the differences between a $ogical% !xtended% and Physical Data +odel.
Identify the difference between a $ogical% !xtended% and Physical Data +odel.
+odule /- Data Distribution% 0ashing% and Index "ccess
Describe the data distribution form and method.
Describe 0ashing.
Describe Primary Index hash mapping.
Describe the reconfiguration process.
Describe a #loc1 $ayout.
Describe File 'ystem ead "ccess.
+odule 2- File 'ystem Writes
Describe File 'ystem Write "ccess.
Describe what happens when Teradata inserts a new row into a table.
Describe the impact of row inserts on bloc1 sizes.
Describe how fragmentation affects performance.
+odule 3- "nalyze Primary Index Criteria
Identify Primary Index choice criteria.
Describe uni*ueness and how it affects space utilization.
!xplain row access% selection% and selecti&ity.
Choose between single and multiple4column Primary Indexes.
+odule 5- Partitioned Primary Indexes
Describe the components that comprise a ow 6ey.
$ist two ad&antages of implementing a PPI.
$ist two potential disad&antages of using a PPI.
Create tables with partitioned primary indexes.
7se the P"TITI.8 1ey word to display partition information.
D"9 2 -
+odule :- 'econdary Index 7sage
Describe 7'I and 87'I implementations.
Describe dual 87'I access.
Describe 87'I bit mapping.
!xplain 87'I and "ggregate processing.
Compare 87'I &s. full table scan ;FT'<.
+odule =- "nalyze 'econdary Index Criteria
Choose columns as candidate 'econdary Indexes.
Describe Composite 'econdary Indexes.
"nalyze Change ating and >alue "ccess.
+odule ?- "ccess Considerations and Constraints
"nalyze .ptimizer "ccess scenarios.
!xplain partial &alue searches and data con&ersions.
Identify the effects of conflicting data types.
Determine the cost of [email protected].
Identify column le&el attributes and constraints.
Identify table le&el attributes and constraints.
"dd% modify and drop constraints from tables.
!xplain how the Identity column allocates new numbers.
+odule A- eferential Integrity
'pecify how is the reference index subtable hash distributedB
'pecify how a reference index row can be mar1ed as Cin&alidD.
+odule ,E- 'izing
Determine column sizing re*uirements based on chosen data type.
Determine physical table row size.
Determine table sizing re*uirements &ia estimates and empirical e&idence.
Determine index sizing re*uirements &ia estimates and empirical e&idence.
+odule ,,- ')$ Parser
Describe internal% channel% and $"8 parcels.
!xplain software cache functionality.
$ist and identify the function of the main components ;phases< of the parser.
Describe the functionality of e*uest4To4'teps cache.
+odule ,/- .ptimizer and Collecting 'tatistics
!xplain how the .ptimizer ac*uires statistics.
Describe dynamic sampling.
'tate the method for &iewing statistics.
Identify the ways to !FP$"I8 +acros.
Describe how the Teradata 'tatistics Wizard can be used to collect or re4collect
statistics.
+odule ,/./- Compiling 'tatistics
'tatistics optimizer statistics
andom amp sampling
Dynamic sampling
Collect statistics syntax
Collecting statistics
>iewing statistics
'tatistics and une&en nusis
7sing the sample feature
'tatistics and ppi tables
'tatistics summary
e&iew *uestions
$ab
+odule ,2- The !FP$"I8 Facility
Describe the !xplain Facility.
Define !xplain terminology.
Describe the !FP$"I8 output of a C!"T! T"#$! statement.
+atch new !FP$"I8 terms ;with PPI< to a definition.
7se !FP$"I8 to determine the number of partitions used in &arious access plans.
DAY 4
+odule ,5- Goin Processing "nalysis
Identify and describe different 1inds of (oin plans.
Identify (oin plan distribution strategies.
Describe different types of (oin strategies that may be used with PPI tables.
+odule ,:- !xplains of Goins and Index Choices
Interpret the !FP$"I8' of 8ested% +erge% and Product Goins.
De&elop 7nion solutions.
!xplain the importance of Goin "ccesses in Index selection.
7se the Index choice guidelines.
7se the Teradata Index Wizard to analyze secondary index selections
+odule ,=- "dditional Index Choices
$ist the three types of Goin Indexes.
Describe the situations where a Goin Index can impro&e performance.
Describe how to create a Csparse indexD.
Describe the difference between a single4table Goin Index and a 0ash Index.

You might also like