Constructing the OLAP Cube from Relational Databases / Flat Files

摘要

Most of us understand spreadsheets and table data. Few can understand databases but only very few can really process the data into meaningful information. Most companies professionals acquaint themselves with pivot tables which summarise data and provide the performance figures at a glance. But for advanced analysis of data, professionals are hired to run the data marts and warehouses. For faster results data in warehouses is often aggregated and presummarised and so called OLAP Cubes are constructed. The start point is usually a flat database(s) or relational databases which are usually designed from the perspective of transactional data. For any OLAP based work, it is said that 90% of efforts are made in construction of the Cube. This paper is written with the perspective of easy understanding of OLAP and allied concepts. Two Sample databases have been referred; one the Food Mart sample database supplied as part of SQL Analysis Services and second is the Global Terror Data (GTD) available on University of Maryland website as a flat file MS Excel Sheet.