Integration of relations with conflicting schema structures in heterogeneous database systems

Abstract

Schema integration is an important discipline for constructing a heterogeneous database system, since there exists various semantic discrepancies among local schemas. However, prior works rarely address the problems of integrating local relations with conflicts of different schema structures. In this paper, we propose an approach for the integration of local relations with such conflicts. In our approach, we define some auxiliary data called metadata for local schemas. The relations can be classified into some classes according to the metadata. The local relations with conflicts of different schema structures are respectively transformed to new relations with the same structure. Finally, we integrate these new relations. We also explore the query decomposition process for global queries. A global query is decomposed into many local queries according to the local schemas' structures. The results of the local queries will be sent to the global site. Finally, the global site integrates the local results into a new one and poses the global query on the integrated result again to get the final result. Since this process filters out unnecessary data in each local site, it can also be regarded as an important work for global query optimization.