Mapping generation for XML data sources: a general framework

The inter-operability of multiple autonomous and heterogeneous data sources is an important issue in many applications such as mediation systems, datawarehouses, or web-based systems. These systems provide a view, called a target schema, on the top of the data sources. Mappings are defined for describing the way instances of the target schema are derived from instances of the data sources. The generation of such mappings is a difficult problem, especially when the target schema and the source schemas are in XML format. In this paper, we propose a framework to automatically find the mappings for a target schema given the source schemas and a set of semantic correspondences. In our framework, the target schema is decomposed into subtrees. Mappings are first determined for each subtree, and then combined to generate the mapping for the whole target schema. The generated mappings are expressed in a standard language, such as XQuery or XSLT.