Reformulation of XML Queries and Constraints

We state and solve the query reformulation problem for XML publishing in a general setting that allows mixed (XML and relational) storage for the proprietary data and exploits redundancies (materialized views, indexes and caches) to enhance performance. The correspondence between published and proprietary schemas is specified by views in both directions, and the same algorithm performs rewriting-with-views, compositionwith-views, or the combined effect of both, unifying the Global-As-View and Local-As-View approaches to data integration. We prove a completeness theorem which guarantees that under certain conditions, our algorithm will and a minimal reformulation if one exists. Moreover, we identify conditions when this algorithm achieves optimal complexity bounds. We solve the reformulation problem for constraints by exploiting a reduction to the problem of query reformulation. Comments Postprint version. Published in Lecture Notes in Computer Science, Volume 2572, Proceedings of the 9th International Conference on Database Theory 2003, (ICDT 2003), pages 225-241. Publisher URL: http://springerlink.metapress.com/link.asp?id=08u0mdgb6u0g34y5 This conference paper is available at ScholarlyCommons: http://repository.upenn.edu/cis_papers/119 Reformulation of XML Queries and Constraints Alin Deutsch and Val Tannen 1 UC San Diego, [email protected] 2 University of Pennsylvania, [email protected] Abstract. We state and solve the query reformulation problem for XML publishing in a general setting that allows mixed (XML and relational) storage for the proprietary data and exploits redundancies (materialWe state and solve the query reformulation problem for XML publishing in a general setting that allows mixed (XML and relational) storage for the proprietary data and exploits redundancies (materialized views, indexes and caches) to enhance performance. The correspondence between published and proprietary schemas is specified by views in both directions, and the same algorithm performs rewriting-with-views, composition-with-views, or the combined effect of both, unifying the Global-As-View and Local-As-View approaches to data integration. We prove a completeness theorem which guarantees that under certain conditions, our algorithm will find a minimal reformulation if one exists. Moreover, we identify conditions when this algorithm achieves optimal complexity bounds. We solve the reformulation problem for constraints by exploiting a reduction to the problem of query reformulation.