A New Framework for Designing Schema Mappings
One of the fundamental tasks in information integration is to specify the relationships, called schema mappings, between database schemas. Schema mappings specify how data structured under a source schema is to be transformed into data structured under a target schema. The design of schema mappings is usually a non-trivial and time-intensive process and the task of designing schema mappings is exacerbated by the fact that schemas that occur in real life tend to be large and heterogeneous. Traditional approaches for designing schema mappings are either manual or performed through a user interface from which a schema mapping is interpreted from correspondences between attributes of the source and target schemas. These correspondences are either specified by the user or automatically derived by applying schema matching on the two schemas.
In this paper, we examine an alternative approach that allows a user to follow the “divide-design-merge” paradigm for specifying a schema mapping. The user can choose to independently design schema mappings for smaller portions of the source and target schema. Afterwards, the user can interact with the system to refine and further design schema mappings through the use of data examples. Finally, in the merge phase, a global schema mapping is generated through the correlation of the individual schema mappings.
KeywordsSchema mappings data examples merge
Unable to display preview. Download preview PDF.