Designing relational data warehouses through schema-transformation primitives

Abstract

A Data Warehouse (DW) is a database that stores information oriented to satisfy decision-making request. It is a database with some particular features concerning the data it contains and its utilisation. The features od DWs cause the DW design process and strategies to be different from the ones for OLTP Systems. We address the DW design problem through a schema transformation approach. We propose a set of schema transformation primitives, which are high-level operations that transform relational sub-schemas into other relational sub-chemas. We also provide some tools that can help in DW design process: (a) the design trace, (b) a set of DW schema invariants, (c) a set of rules that specify how to correct schema-strategies dor designing the DW through application of primitives.