Rewriting queries for XML integration systems

Abstract

A data integration system typically creates a target XML schema to represent an application domain and source schemas are mapped to the target schema. A user poses a query over the target schema, and the system rewrites the query into a set of queries over the data sources. Existing algorithms generate a set of static rules based on the target schema and mappings, and rewrite the target query using these rules. We design a flexible and dynamic approach that rewrites XML queries directly based on the mappings between the target and source schemas. Theoretical analysis and experiments on both synthetic and real-world datasets indicate that the proposed approach is efficient and scalable. © Springer-Verlag Berlin Heidelberg 2006.