Abstract
In our research work, we attempt to bridge two fields once closely interconnected: Information Preservation and Data Integration. This relative split is due both part to the growth of the web and its associated challenges and the lack of proper formalism and tools for lossless schema transformation. To answer that concern, our approach is to propose a catalogue of lossless schema transformation templates, called transformation patterns, based on operations commonly used in data integration. As some of these operations include the promotion and/or demotion of metadata in a relational schema, we introduce an extension of transformation patterns by incorporating a dynamic aspect. We also want to deal with the leftover operations, those that lose information, and turn them into lossy transformation patterns. Finally, we present the early implementation of a schema transformation framework using transformation patterns.