This document compares and contrasts dimensional modeling and E-R modeling. It provides examples of a star schema and snowflake schema dimensional model. It notes that dimensional modeling supports OLAP and online analytical processing while E-R modeling supports OLTP. Dimensional modeling uses a denormalized structure while E-R modeling is normalized. Dimensional modeling is more robust and understandable compared to E-R modeling.

1. Comparative Study of E-R Modeling and Dimensional Modeling
PRN
NAME
12030121020
Shivam Tyagi
12030121026
Aniket Srivastava
12030121028
Amit Shivsharan
12030121030
Karuna Kak
BCA (2012-15)
Division A

2. Entity Relationship Modeling
•Theentity-relationship(E-R)datamodelisbasedonaperceptionofarealworldthatconsistsofasetofbasicobjectscalledentities,andofrelationshipsamongtheseobjects.
•StepsinE-RModeling:
Identify the entitiesLook for general Nouns in requirement specification document, which are of business interest to users.
IdentifythekeyattributesforeveryentityIdentifytheattributeorsetofattributeswhichcanidentifyinstanceofentityuniquely
Find the associations between entitiesIdentify the natural relationship and their cardinalities between the entities.
Complete E-R Diagram Draw complete ER diagram with all attributes including primary key .
2

3. 3
Symbol
Description

4. E-R Modeling Example
4

5. Star schema
•Starschemaisadimensionmodelingtechnique.
•ThedimensiontablearerelatedtoeachotherwithM:Mrelationshipandaredenormalized.
•ThefacttableisnormalisedbecauseithasI:Mrelationshipwithotherdimensions.
•Example:Items,Branch,Location,andTimeareDimensionTableswhicharedenormalized.Totalsaleswithtime_key,item_key, branch_keyandlocation_key(primarykeysofalldimensiontables)istheFactTablewhichisnormalized.
5

6. 6
Star Schema Example

7. Snowflake schema
•Thesnowflakeschemaisadimensionmodelingtechnique.
•"Snowflaking"isamethodofnormalisingthedimensiontablesinastarschema.Dimensiontablesgetpartiallynormalised.
•Theprinciplebehindsnowflakingisnormalisationofthedimensiontablesbyremovinglowcardinalityattributesandformingseparatetables.
•Whenwedonotaccessaparticulartableorwedon’tretrieveanythingfromthetablefrequently,webreakthetablewith1:Mrelationship.
•Example:Items,Branch,Location,andTimeareDimensionTableswhicharedenormalized.Totalsaleswithtime_key,item_key,branch_keyandlocation_key(primarykeysofalldimensiontables)istheFactTablewhichisnormalized.Alsoitemsdimensionhadcityfieldsandlocationdimensionhadacityfieldswhichwereretrievedveryinfrequentlythereforetheyarebreakdownintoseparatedimensiontables.
7

8. 8
Snowflake Schema Example

9. Comparative Study of ER Modeling and Dimensional Modeling
E-R Modeling
Dimensional Modeling
SupportsOLTP(OnlineTransactionProcessing)
SupportsOLAP(OnlineAnalyticalProcessing)
Entitiesarelinkedwithaseriesofjoins.
Entitiesarelinkedwithaseriesofjoins.
Normalized
Denormalized
Removesredundancy.
Permitsredundancy.
Ifmodelismodified,theapplicationaremodified.
Itisextensibletoaccommodateunexpectednewdataelementsandnewdesigndecision.
Itisvariableinstructureandveryvulnerabletochangesintheuser’squeryinghabits.
Itisrobust.Thedimensionalmodeldesigncanbedoneindependentofexpectedquerypattern.
Themodelforenterpriseisveryhardforpeopletovisualizeandkeepintheirheads.
Thismodeliseasyandunderstandable.