A Multidimensional Data Model Supporting Complicated Aggregations

Abstract

A data cube is a collection of multidimensional data not only of one level, nor of all levels by one aggregation function, but of all levels by all aggregation functions admissible in an OLAP system. We propose a multidimensional data model that produces such kind of data cubes so that aggregation data of different levels can be kept consistent. The multidimensional data of different levels in a data cube are organized with so called aggregation nodes. We also define the operations on a data cube and illustrate the expressive power of the model by expressing analysis queries.