Coffing Dw Physical-Table of Contents

7/30/2019 Coffing Dw Physical-Table of Contents

1/9

Coffing Data WarehousingEducation Outline02/17/05

TERADATA EDUCATIONOUTLINE

Coffing Data Warehousing has provided quality Teradata education, productsand

services for over a decade. We offer customized solutions to maximize yourwarehouse.

Toll Free: 1-877-TERADATBusiness Phone: 1-937-855-4838

Email: mailto:[email protected]: http://www.CoffingDW.com

In addition to the course material listed in this outline, we also offer Teradata classes in TeradataBasics, Implementation, SQL, Database Administration, Design and Utilities.Please contact us so we can customize a course to fit your specific needs.

2006 Coffing Data Warehousing All rights reserved. Confidential. 1
http://www.coffingdw.com/http://www.coffingdw.com/


2/9


PURPOSE

Coffing Data Warehousing has been providing quality Teradata education for over adecade. We offer customized courses to maximize the effectiveness of each class.

The purpose of this proposal is to build a lasting relationship with your company. To thisend, we have combined our comprehensive Teradata education services in a uniquepackage that we feel best suits the diverse needs of your company while offering ourhigh quality product at competitive pricing.

Coffing Data Warehousing is excited to offer you, our preferred partner, an innovativenew way to look at training at the CoffingDW Teradata University (CDW-TU). Thisapproach provides the ability to maximize learning potential. Our goal is to make youremployees the most educated data warehouse experts in the industry.

CURRICULUM:

Coffing Data Warehousing will provide an experienced and highly qualified resource todeliver this customized educational seminar on the following topic(s):

Teradata Education

Teradata Physical Implementation

COURSE DESCRIPTION

COURSEPREREQUISITES There is no prerequisite for this course.

COURSE This course is designed to be highlyDuration/Format interactive with the audience.

COURSEAUDIENCE The audience will consist of a mix ofbeginning,intermediate and advanced

Teradata users.

OBJECTIVES This course is designed to provide in-depthknowledge of Teradata PhysicalImplementation.



3/9


Tera-Tom on Teradata Physical Implementationfor V2R6

Chapter 1 The Rules of Data Warehousing

Teradata CertificationA Logical View of the Teradata Architecture

The Parsing Engine (PE)The Access Module Processors (AMPs)The BYNETA Visual for Data Layout

Teradata Cabinets, Nodes, Vprocs, and Disks

Chapter 2 Data Distr ibution Explained

Rows and ColumnsThe Primary IndexThe Two Types of Primary IndexesUnique Primary Index (UPI)

Turning the Primary Index Value into the Row HashThe Row Hash Value determines the Rows DestinationThe Row is Delivered to the Proper AMP

The AMP will add a Uniqueness ValueAn Example of an UPI TableAn Example of a NUPI TableHow Teradata Retrieves RowsRow DistributionA Visual for Data Layout

Teradata accesses data in three waysData Layout Summary

Chapter 3 V2R5 Parti tion Primary Indexes

V2R4 ExampleV2R5 PartitioningPartitioning doesnt have to be part of the Primary IndexPartition Elimination can avoid Full Table Scans

The Bad NEWS about Partitioning on a column that is not part of thePrimary Index



4/9


Two ways to handle Partitioning on a column that is not part of thePrimary IndexPartitioning with CASE_NPartitioning with RANGE_N

NO CASE, NO RANGE, or UNKNOWNPartitioning and J oins

Chapter 4 Teradata Engine Under the Hood

Full Cylinder ReadTable HeaderEach Table is given a Table IDHow Data Blocks are Dynamically BuiltData BlocksHow Teradata Finds a Row of Data

The Master IndexThe Cylinder IndexCylinder Index ChangesHow Teradata Writes to an AMPWriting to Data Blocks of Equal LengthWhen a Data Block is not Big Enough for a WriteHow Teradata Allocates BlocksBlock and Row Definitions

Large Row versus Oversized RowDefragmentationWhen a Cylinder becomes FullA Node and its Memory Allocations

Chapter 5 The Extended Logical Data Model

The Application Development Life CycleAsking the Right QuestionsLogical Data ModelPrimary KeysForeign KeysNormalizationExtended Logical Data Model

The End Goal of the ELDM is to build Table TemplatesColumn ACCESS in the WHERE Clause



5/9


Data DemographicsDistinct Values

Typical Rows Per ValueMaximum Rows NULL

Change RatingExtended Logical Data Model Template

Chapter 6 The Phys ical Data Model

Step 1 Look at DistributionStep 2 Eliminate based on Change RatingStep 3 NUSI Elimination via Value Access FrequencyStep 4 Pick the Primary IndexPrimary Index FactorsWhy J oin Access Frequency is Top Priority?Why Value Access Frequency is Second Priority?What have we learned about picking the Primary Index?Step 5 Pick Secondary IndexesUSI to eliminate Duplicate Row CheckingNUSI considerationsMulti-Column NUSI Columns used as a Covered QueryValue-Ordered NUSIsA Strongly Vs Weakly Selective NUSI

A formula for calculating a strongly selective NUSITypical Row SizeTypical Block Size

Chapter 7 Denormalization

Derived DataStoring AggregatesPreJ oining TablesRepeating GroupsHorizontal PartitioningVertical PartitioningCovered QuerySingle-Table J oin IndexesMulti-Table J oin Indexes

Temporary Tables



6/9


Derived TablesVolatile Temporary TablesGlobal Temporary Tables

Chapter 8 Secondary Indexes

Unique Secondary Index (USI)USI Subtable ExampleHow Teradata retrieves an USI queryNUSI Subtable ExampleHow Teradata retrieves a NUSI queryValue-Ordered NUSIHow Teradata retrieves a Value-Ordered NUSI queryNUSI BitmappingPrototyping indexes with EXPLAINSecondary Index Summary

Chapter 9 Join Strategies

A J oin in Simple TermsThe key things to know about Teradata and J oinsMerge J oin Strategies

J oins need the joined rows to be on the same AMP

Another Great J oin PictureJ oining Tables with matching rows on different AMPsJ oining Tables with matching rows on different AMPsRedistributing a Table for J oin PurposesBig Table Small Table J oin StrategyBig Table Small Table DuplicationNested J oinHash J oinExclusion J oinProduct J oinsCartesian Product J oin

Chapter 10 Join Indexes

Three basic types of J oin IndexesJ oin Index Fundamentals



7/9


J oin Indexes versus other objectsMulti-Table J oin IndexSingle-Table J oin IndexesAggregate J oin Index

Sparse IndexSparse Index PictureGlobal J oin IndexGlobal J oin Index PictureGlobal J oin Index Multi-Table J oin BackHash IndexesHash Indexes vs. Single-Table J oin Indexes

Chapter 11 Explains

The Teradata Optimizer knows how to Explain in DetailRow Estimate Confidence LevelsExplain TerminologyVisual Explain

Chapter 12 The Parsing Engine in Detail

The Parsing Engine (PE) goes through Six StepsEach PE has a Plan Library called RTS Cache

The Parsing Engine has Data Dictionary CacheWhy the PE loves the Macro

The Parsing Engine in DetailThe Parsing Engine Knows All

Chapter 13 Understanding Views and Macros

ViewsCreating Simple VIEWs and VIEWs that J oin TablesHow to Change a VIEW Using REPLACEHow to Drop a VIEWView Aggregation and Nested ViewsAll About MacrosMacros that Use ParametersChanging a MACRO Using REPLACEHow to Drop a MACRO



8/9


Chapter 14 - Locks

Teradata has Four locks at Three levels

Locks and their compatibilityHow Teradata Locks Objects

Teradata Locks First Come First ServeLocking Queue Example 2Locking Queue Example 3Locking Modifier

The NOWAIT Option

Chapter 15 - Collect Statist ics

Dynamic AMP SamplingHow Collect Statistics WorksSample StatisticsSample StatisticsWhat You Should COLLECT STATISTICS OnCOLLECT STATISTICS DETAILED SYNTAXCOLLECT STATISTICS Examples

Chapter 16 - MISC

Identity ColumnsIdentity Columns ExampleLIKE ClauseSUBSTRING and SUBSTR FunctionsReferential IntegritySoft Referential IntegrityMaterialized ViewsCompressionImplementing CompressionHow Compression WorksCreating a Table With DATABLOCKSIZE

The ALTER TABLE CommandIndex WizardIndex Wizard



9/9



Documents

Coffing Dw Physical-Table of Contents