4 Conclusions and Future Work 3.1 Selection of Models 2.2 Complex Structured Dimensions 2.3 Complex Structured Cell Values (measures) 2.4 Typical Queries Finding Your Way through Multidimensional Data Models 1

It is shown that every MDD table can be represented by a classical relation and vice-versa which allows to base the approach on the relational algebra. The algebra of [GL97] introduces the classical operators The classical operators are defined as the tabular (i.e. multidimensional) representation of the operator applied to the relational representation of a given table instance. This allows a very simple definition of the operators (and leads to an expressive power comparable to the relational algebra) because it can be based on the relational algebra. Derived measures must be materialized and stored in the fact table rm. Dimension hierarchies are modeled in the attribute names, i.e. every hierarchy is stored as separate attribute of a dimension. The same applies to multiple hierachies on a dimension. Complex measures can be modeled as additional attributes in the fact table rm. Example: First, we define a four-dimensional table schema vehicle_repair=<{time, garage, vehicle},