Providing Semantically Equivalent, Complete Views for Multilingual Access to Integrated Data

Today, there are several approaches supporting an integrated processing of XML and (object-)relational data sources. However, each of these approaches favors a single data model for integration purposes and, thus, provides just a single query interface, usually either SQL or XQuery. In order to be more flexibly applicable, our SQXML integration system provides both query languages at its interface. The challenge here is to ensure that the complete information resulting from the data integration process is accessible via both query languages. In this paper we prove that our SQXML integration approach meets this objective. For this purpose we examine the schema transformation, mapping and integration steps performed in the SQXML integration and view generation process in order to show that the resulting SQL and XML views are semantically equivalent in the sense that each encompasses the complete information resulting from schema integration.