Time aspects in SAusiness Information Warehouse CE 3, 3-07-27 Dr. Michael Hahne cundus AG branch manager michael.hahne@cundus.de www.cundus.de Michael Hahne(3)
Contents Multidimensional data structures Time aspects in a Data Warehouse Business Information Warehouse Enhanced Star Schema of SAP Variants for modeling hierarchical structures in dimensions Slowly changing dimensions in BW CE 3, 3-07-27, Michael Hahne 2
Contents Multidimensional data structures Time aspects in a Data Warehouse Business Information Warehouse Enhanced Star Schema of SAP Variants for modeling hierarchical structures in dimensions Slowly changing dimensions in BW CE 3, 3-07-27, Michael Hahne 3
dimensions and balanced hierarchical structures derived resp. consolidated elements dimension members hierarchy level resp. consolidation level Base elements resp. independent elements granularity CE 3, 3-07-27, Michael Hahne 4
unbalanced dimension structure derived resp. consolidated elements dimension members hierarchy level resp. consolidation level Base elements resp. independent elements granularity CE 3, 3-07-27, Michael Hahne 5
Contents Multidimensional data structures Time aspects in a Data Warehouse Business Information Warehouse Enhanced Star Schema of SAP Variants for modeling hierarchical structures in dimensions Slowly changing dimensions in BW CE 3, 3-07-27, Michael Hahne 6
Solutions for structural changes in dimensions Modifying historical data to match new structures pros: small data sets; data structures remain clear cons: old structures are lost even if users want to analyze data with old structures Save historical data separately in addition to the complete data matching new structures pros: Old reports can be reproduced cons: Higher volume of data; actualizing is complicated and costly if new data should be analyzed according to old structures; confusing for end-users Build parallel hierarchies with old and new structure pros: All data can be analyzed with any structure cons: dimension structure is very confusing Temporal databases - time stamps pros: All data can be analyzed with any structure cons: poor performance; concepts not fully developed CE 3, 3-07-27, Michael Hahne 7
Contents Multidimensional data structures Time aspects in a Data Warehouse Business Information Warehouse Enhanced Star Schema of SAP Variants for modeling hierarchical structures in dimensions Slowly changing dimensions in BW CE 3, 3-07-27, Michael Hahne 8
Architecture SAW 3.0 Third-Party- Tools Business Explorer BAPI ODBO XML Administrator Workbench BW-Server OLAP-Processor Info-Cubes Administration Scheduling Metadata- Manager Data-Manager ODS Monitoring Staging-Area PSA BAPI FILE XML XML-Daten flat files Non-SAP SAP R/2 SAP R/3 SAW CE 3, 3-07-27, Michael Hahne 9
Contents Multidimensional data structures Time aspects in a Data Warehouse Business Information Warehouse Enhanced Star Schema of SAP Variants for modeling hierarchical structures in dimensions Slowly changing dimensions in BW CE 3, 3-07-27, Michael Hahne 10
Concept of master data in BW attributes customer: customer group texts customer : language, caption hierarchies customer : customer hierarchy dimension customer Info-Cube dimension time attributes time: legal holiday fact table of the Info-Cube dimension table customer dimension table time texts time : language, caption dimension table product hierarchies time : calender hierarchy attributes product: product group texts product : language, caption hierarchies product : product hierarchy dimension product CE 3, 3-07-27, Michael Hahne 11
SID-tables connecting the master data tables C P T sales revenue... 20 0 50 2500 fact table Info-Cube dimension table time dimension table product C SID-customer SID-trust SID-branch SID-headquarter 522 170 dimension table customer SID-customer customer 522 3417226 SID-region region 23 S5 SID-region 23 SID-Table for attributes SID-Table for attributes customer text 3417226 Billy Bikes Text-Tables for descriptions in different languages region text S5 London South5 SID-trust trust 170 12769 SID-Table for attributes master data CE 3, 3-07-27, Michael Hahne 12
different master data tables K SID-cust. SID-trust SID-branch SID- headq. 522 170 Dimension Table customer S SID-cust customer 522 3417226 SID-Table BIC/SKunde Standard SID-Table P Q customer Customer name 3417226 Lampen Müller customer 3417226 DateFrom... DateTo... region S5 Master Data Table BIC/PKunde for not time-dependant Display Attributes Master Data Table BIC/QKunde for time-dependant Display Attributes X SID-cust customer 522 3417226 SID-region 23 SID-Table BIC/XKunde for not time-dependant Navigational Attributes Y SID-cust customer DateFrom 522 3417226... DateTo... SID-class 12 SID-Table BIC/YKunde for time-dependant Navigational Attributes CE 3, 3-07-27, Michael Hahne 13
Contents Multidimensional data structures Time aspects in a Data Warehouse Business Information Warehouse Enhanced Star Schema of SAP Variants for modeling hierarchical structures in dimensions Slowly changing dimensions in BW CE 3, 3-07-27, Michael Hahne 14
Modeling Hierarchical structures in BW Model region and country country as Characteristic? In Dimension Table region as Navigational Attribute In Master Data Table customer as Hierarchy In Hierarchy-Table CE 3, 3-07-27, Michael Hahne 15
Hierarchical structures between characteristics country region customer dimension table SIDs characteristic customer characteristic region characteristic country CE 3, 3-07-27, Michael Hahne 16
Hierarchical structures based upon navigational attributes region country customer SID dimension table characteristic customer SIDs master data table attributes attribute region attribute country CE 3, 3-07-27, Michael Hahne 17
Hierarchies in BW edges dimension table customer SID characteristic customer child parent master data table hierarchy (inclusion table) text node CE 3, 3-07-27, Michael Hahne 18
Contents Multidimensional data structures Time aspects in a Data Warehouse Business Information Warehouse Enhanced Star Schema of SAP Variants for modeling hierarchical structures in dimensions Slowly changing dimensions in BW CE 3, 3-07-27, Michael Hahne 19
Slowly Changing Dimensions - Example Product dimension in Product Product Group Fact Table Product P E Product dimension in Product Group (changed) (new) Product P E Time Revenue CE 3, 3-07-27, Michael Hahne 20
time scenarios possible reporting demands Report using today s constellation Product Group Revenue Revenue 300 400 Report using constellation of Product Group Revenue Revenue Reporting historical truth Product Group Revenue Revenue 400 Reporting comparable results Product Group Revenue Revenue CE 3, 3-07-27, Michael Hahne 21
Scenario I : report data with today s constellation Product dimension in Fact Table Product Product Group Product Time Revenue P E (changed) (new) P E Report using today s constellation Product Group Revenue Revenue 300 400 CE 3, 3-07-27, Michael Hahne 22
Scenario I : Solutions in BW Product Group as Navigational Attribute of the Characteristic Product join from Product Group SID-Table to Product Attribute SID-Table to Dimension Table to Fact Table Modelling the relation Product Group -> Product in an External Hierarchy join from Product Hierarchy Table to Dimenson Table to Fact Table CE 3, 3-07-27, Michael Hahne 23
Scenario II : report data with yesterday s constellation Product dimension in Product Product Group Product P E Report using constellation of Product Group Revenue Revenue Missing Product P E Fact Table Time Revenue CE 3, 3-07-27, Michael Hahne 26
Scenario II : Solutions in BW Product Group as time-dependant Navigational Attribute of the Characteristic Product join from Product Group SID-Table to Product time-dependant Attribute SID-Table with Query-Keydate to Dimension Table to Fact Table Modelling the relation Product Group -> Product in an External Hierarchy with versioning, time-dependant hierarchy or even time-dependant structure join from Product Hierarchy Table with Query Keydate to Dimenson Table to Fact Table CE 3, 3-07-27, Michael Hahne 27
Scenario III : Reporting historical truth Product dimension in Product Product Group Product dimension in Product P E Product Group (changed) (new) Product P E Fact Table Time Revenue Reporting historical truth Product Group Revenue Revenue 400 CE 3, 3-07-27, Michael Hahne 30
Scenario III : Solution in BW Model Product Group and Product as Characteristics in the same dimension join from Product Group SID-Table to Dimension Table to Fact Table CE 3, 3-07-27, Michael Hahne 31
Scenario IV: reporting comparable results Product dimension in Product Product Group Product dimension in Product P E Product Group (changed) (new) Product P E Fact Table Time Revenue Reporting comparable results Product Group Revenue Revenue CE 3, 3-07-27, Michael Hahne 33
Scenario IV : Solution in BW Product Group as time-dependant Navigational Attribute of the Characteristic Product with additional time-dependant attributes join from Product Group SID-Table to Product time-dependant Attribute SID-Table with Query-Keydate to Dimension Table to Fact Table to ensure the validity of structure in all reported periods we need additional time-dependant attributes FROM and TO, because the system attributes can t be used in a query to filter Query with Keydate in the reported time interval, FROM le lower bound of reported period, TO ge upper bound of reported period CE 3, 3-07-27, Michael Hahne 34
contact Dr. Michael Hahne cundus AG branch manager michael.hahne@cundus.de www.cundus.de Michael Hahne(3)