SAP BW Implementation at JoAnn Stores Inc. Session Code 3309 Craig Eick JoAnn Stores Inc.
Agenda JoAnn Stores Inc. Overview Project Overview and Background Data Required Model Design InfoCube Partitioning (Fact Table) Extracting and Updating of Data Aggregation Compression Validation
Agenda JoAnn Stores Inc. Overview Project Overview and Background Data Required Model Design InfoCube Partitioning (Fact Table) Extracting and Updating of Data Aggregation Compression Validation
JoAnn Stores Overview (I) Craft and Fabric Retailer with over 940 stores in 49 states. –96 Etc. Stores (Stand alone, 40,000 sq. ft and above). –849 Traditional Stores (Strip plazas). 318,000 articles with > 95,000 active. Serviced by > 3500 vendors, with 340 being import vendors.
JoAnn Stores Overview (II) BW Application Server –RS/6000 (p690) 32 gig RAM 16 CPU’s AIX L V5.2 –500 GiG of space allocated to production. DB2 Server –z/OS Mainframe –DB2 Version 7.1 –Roughly a 3 terabyte Farm for DB2
Agenda JoAnn Stores Inc. Overview Project Overview and Background Data Required Model Design InfoCube Partitioning (Fact Table) Extracting and Updating of Data Aggregation Compression Validation
Project Overview and Background Will function as a means to eliminate many of the Access databases that exist in different functional areas within the Company To enable some of the scorecarding requirements of different functional areas so that accountability may be raised To improve new reporting turnaround that now is done via RIS InfoStructures and via ABAP reports. Port over 2 of our summary level RIS InfoStructures
Agenda JoAnn Stores Inc. Overview Project Overview and Background Data Required Model Design InfoCube Partitioning (Fact Table) Extracting and Updating of Data Aggregation Compression Validation
Data Required Article Characteristics Attributes that are in a select number of Access Databases Attributes that are on key ABAP reports Total of 51 attributes selected 3 Time dependent attributes Article based on Site Total of 13 attributes selected 4 Time dependent attributes
Data Required
Site Characteristics Total of 38 attributes selected 4 Time dependent attributes
Data Required
Vendor Characteristics Total of 16 attributes selected
Data Required
Key Figures Taken from existing RIS InfoStructures Total of 50 stored key figures Total of 37 Calculated key figures
Data Required
Agenda JoAnn Stores Inc. Overview Project Overview and Background Data Required Model Design InfoCube Partitioning (Fact Table) Extracting and Updating of Data Aggregation Compression Validation
Model Design 2 InfoCube Design (based on 2 summary level RIS InfoStructures InfoCube #1 - Article, Store Type, Week InfoCube #2 - Merchandise Category, Site, Week
Model Design
Agenda JoAnn Stores Inc. Overview Project Overview and Background Data Required Model Design InfoCube Partitioning (Fact Table) Extracting and Updating of Data Aggregation Compression Validation
InfoCube Partitioning (Fact Table) Fact Table partitioned by JoAnn’s Fiscal Year InfoCube for 2002, for 2003, for 2004, etc Article Site Vendor Time MultiCube
InfoCube Partitioning (Fact Table) Fact Table Partitioned by Key Figure Set Open Orders All other Key Figures (Sales, Inventory, Receipts, etc.) 2004 Open Orders Sales, Inventory, etc. Article Site Vendor Time MultiCube
InfoCube Partitioning (Fact Table) Fact Table Partitioned by Logical Unit Vendor Number on 1 InfoCube, Site on the Other 7 range split On Sales, Inventory, etc. cubes 2004 Sales, Inventory, etc. Article Site Vendor Time MultiCube Vendor range 1, 2, 3 Vendor range 4, 5, 6 Vendor range 7
InfoCube Partitioning (Fact Table) Article Site Vendor Time MultiCube Sales, Inventory, etc. Open Order Open Order Open Order Article/Store Type/Week InfoCube
InfoCube Partitioning (Fact Table)
Agenda JoAnn Stores Inc. Overview Project Overview and Background Data Required Model Design InfoCube Partitioning (Fact Table) Extracting and Updating of Data Aggregation Compression Validation
Extracting and Updating of Data Master Data 6 ABAP Extract Programs 42 Flat Files created 42 InfoPackages to update data in BW 3rd Party Scheduler (Tivoli) controls the extracts and update jobs/tasks.
Extracting and Updating of Data Master Data (Continued) Article Delta Logic To reduce # of updates on a daily basis Compare yesterday’s extract file to today’s. Check for any attribute change, if any, article is put in delta file Article and Site Time Dependant Logic Weekly compare to previous runs extract file. If any of the time dependant fields, the output file gets a new record with a time stamp on it.
Extracting and Updating of Data Transactional Data History Extract & Load ABAP programs to extract from our existing RIS InfoStructures 2 + years (24 + fiscal months) flat files per fiscal month Number of files increases because each file contains 400,000 records. Each time a file reaches 400,000 records, a new file is created.
Extracting and Updating of Data Transactional Data (continued) Daily Update Same jobs/programs that perform daily update to RIS InfoStructures, create flat files to go to BW. Extract 52 Wide, creating 52 flat files
Extracting and Updating of Data
Agenda JoAnn Stores Inc. Overview Project Overview and Background Data Required Model Design InfoCube Partitioning (Fact Table) Extracting and Updating of Data Aggregation Compression Validation
Aggregation 2 Aggregates per cube to start with. Company level (highest hierarchy node) Category level (2nd lowest hierarchy node) We will address additional levels of aggregation as new queries are designed.
Aggregation Aggregation Execution 1 Batch job per InfoCube (16 total) 4-5 hour runtime per quarter for the history Loads
Aggregation
Agenda JoAnn Stores Inc. Overview Project Overview and Background Data Required Model Design InfoCube Partitioning (Fact Table) Extracting and Updating of Data Aggregation Compression Validation
Compression
Agenda JoAnn Stores Inc. Overview Project Overview and Background Data Required Model Design InfoCube Partitioning (Fact Table) Extracting and Updating of Data Aggregation Compression Validation
Master Data Initial Validation All done by user via Bex and R/3 downloads. Data Integrity department, segregated master data fields by R/3 table. Loaded both to Access tables, executed queries to find differences.
Validation History Loads ABAP Balancing of Extract files to RIS InfoStructures ABAP Balancing of InfoCubes to Extract files Bex Queries to report InfoCube totals Complete set of User defined Validation Steps Specific Hierarchy nodes To make sure aggregate tables are hit. All the way down to Article and/or Site
Validation
Click to add text Click to add title
Click to add text Click to add title
Thank you for attending! Please remember to complete and return your evaluation form following this session. Session Code: 3309 Craig Eick JoAnn Stores Inc.