Query rewriting with symmetric constraints

Article type: Research Article

Authors: Koch, Christoph

Affiliations: Database and Artificial Intelligence Group, Technische Universität Wien, A‐1040 Vienna, Austria E‐mail: [email protected]

Abstract: We address the problem of answering queries using expressive symmetric inter‐schema constraints which allow to establish mappings between several heterogeneous information systems. This problem is of high relevance to data integration, as symmetric constraints are essential for dealing with true concept mismatch and are generalizations of the kinds of mappings supported by both local‐as‐view and global‐as‐view approaches that were previously studied in the literature. Moreover, the flexibility gained by using such constraints for data integration is essential for virtual enterprise and e‐commerce applications. We first discuss resolution‐based methods for computing maximally contained rewritings and characterize computability aspects. Then we propose an alternative but semantically equivalent perspective based on a generalization of results relating to the database‐theoretic problem of answering queries using views. This leads to a fast query rewriting algorithm based on AI techniques, which has been implemented and experimentally evaluated.

Keywords: Heterogeneous databases, data integration, query rewriting, symmetric constraints, global‐as‐view, local‐as‐view

Journal: AI Communications, vol. 17, no. 2, pp. 41-55, 2004