Data level conflicts resolution for multi-sources heterogeneous databases

Integration of data from multiple, heterogeneous databases is a commonly encountered scenario in information retrieval systems where the user is to be provided with a unified view of information. These data sources may be from different vendors, may comprise different schemas and, could be physically at different locations. A number of techniques have been developed to merge schemas of multiple databases and address the resulting conflicts. While schema integration and resolution of schema conflicts has been an active area of research, relatively lesser attention is paid to the issue of data level conflicts. These conflicts may arise while merging the actual data that is retrieved from multiple databases. A dedicated mechanism needs to be implemented to identify and resolve the conflicts encountered during data integration which makes the subject of our study. This paper identifies few major and frequently occurring data level conflicts which are likely to arise while integrating data. We propose a framework that identifies and resolves these conflicts. The framework is implemented as a software system which we call DHResol. The system evaluated to merge two different databases successfully identified and resolved the discussed conflicts.