Aggregation Everywhere: Data Reduction and Transformation in the Phoenix Data Warehouse

This paper describes the Phoenix system, which loads a data warehouse and then reports against it. Between the raw atomic data of the source system and the business measures presented to users there are many computing environments. Aggregation occurs everywhere: initial bucketing by the natural keys on the mainframe, loading the fact table using a mapping table, maintaining aggregate tables and reporting tables in the data base, in the GUI, in SQL queries issued on behalf of client tools by the web server, and inside the client tools themselves. Producing the desired business measures required writing a complex SQL query, and post-processing an Excel pivot table. Keywords Data Warehouse, OLAP, Aggregation, SQL, Data Lineage

• Type: Conference
• Year: 1999
• Where: DOLAP
• Authors: Steven Tolkin
• Keywords: Business Measures, Data Warehouse, DOLAP 1999, Information Management, Raw Atomic Data