A Conceptual Modelling Perspective for Data Warehouses
Jaroslav Pokorný, Peter Sokolowsky · 1999
The volume of information of the most various types stored electronically in a company is increasing to an ever-greater extent. While in the field of operational systems everything is aimed at achieving the quickest possible throughput, in the dispositive field, questions regarding the total overview or detailed views are of interest. OLAP servers are multidimensionally structured. They are therefore suited for the analysis of multidimensional datastores. The functionality of the OLAP server, such as, creation of forms, drill down, roll up, slice and dice, analysis technology, multidimensional consolidation, etc. demonstrate the advantages of this tool. The analysis of the relevant features of the data warehouse and OLAP is based on both the mainstream literature and on our experience in a two-year project Data Warehouse for Tupperware Inc.The second problem addressed by this paper is the discussion of recent approaches for a proposition some formal definitions of basic constructs used in so called multidimensional modelling which seems to be an important technique for data warehousing and OLAP. It is different from E-R modelling and offers a number of important advantages that the E-R modelling lacks.We show a relationship of E-R modelling to the multidimensional modelling and describe a broad class of multidimensional databases based on so called constellation schemes with explicit hierarchies.