Metadata for Data Warehousing Using Extended Relational Models

This paper introduces different extended relational concepts to model metadata for data warehousing. In the first approach we will use the concept of nested relations to model metadata for multi-dimensional OLAP data. We will discern between the use of the nested relation approach for creating user views which are deduced from a relational model ? which omit the requirement of being in first normal form from the use of nested relations for the conceptual modeling of OLAP. In the second approach we will briefly introduce the use of quotient relations which also inherits the advantage of the original relational model as it was introduced by Codd paired with many advantages for the drill-down and roll-up operations. The powerful partitioning and de-partitioning operators on quotient relations can perform drill-down and roll-up operations in a very convenient way. In the third approach we propose the metadata for multidimensional data using an extended relational model which was introduced by Codd for use in OLAP-modeling. In this model Codd introduced important semantic concepts, such as unique object identifiers, different types of objects (kernel, characteristic and association) and also meta relations, e.g. association graph relation, characteristic graph relation, etc. The general requirements for all three extended relational data models, which could serve as a foundation for multidimensional database systems, are similar to those that made the relational model successful, namely the existence of an implementation independent formalism, the separation of structure and contents, and the existence of declarative query language. All three approaches are compared with the modeling based on the traditional flat relations, which is widely used in the OLAP community, i.e. the modeling of OLAP star and snow-flake schemas. Keywords: data warehouse, OLAP, metadata, extended relational model.