OLAP in Multifunction Multidimensional Database

Abstract

Most models proposed for modeling multidimensional data warehouses consider a same function to determine how measure values are aggregated. We provide a more flexible conceptual model allowing associating each measure with several aggregation functions according to dimensions, hierarchies, and levels of granularity. This article studies the impacts of this model on the multidimensional table (MT) and the OLAP algebra [11]. It shows how the MT can handle several aggregation functions. It also introduces the changes of the internal mechanism of OLAP operators to take into account several aggregation functions especially if these functions are non-commutative.