Using perspective schema and a reference model to design the ETL process
Mar 2011
For many years the database community has been wrestling with the problem of data integration. In short, this problem consists of combining data from different information sources, and providing the user with a unified view of these data. In dealing with the integration problem, the designer usually encounters incompatible data models, characterised by differences in structure and semantics. In this work, we propose a declarative approach based on the creation of a reference model and perspective schemata in order to make the relationship between schemata explicit. The reference model serves as a common semantic meta-model, while perspective schemata defines correspondence between schemata using correspondence assertions. Based on the proposed architecture, an inference mechanism was developed to allow the (semi-) automatic derivation of new mappings from previous ones. A Prolog-based proof-of-concept was implemented in order to demonstrate the feasibility of the proposed approach.