| With the enhancement of social information degree, the form and contents of the information sharing are becoming more and richer. However the heterogeneous among different data sources makes data sharing more difficult. To solve this problem, the key means is effective data integration. The paper mainly studies the integration method of multi-source heterogeneous data. The goal is to provide a uniform query interface to those users who intend to access heterogeneous data source, but not consider the difference among each local data source.At present, among those different integration methods of the heterogeneous data sources, the Mediator method is the main one for its advantages such as low cost, easy to realize, flexible and convenient to add local data source, strong autonomy and so on.At the end of 90's, W3C organization issued XML technology. Its characteristics include the independence of the platform and language, self-descriptiveness, separation of the content and expression, strong scalability and so on. These characteristics make it more suitable to describe any kind of different data model.Based on a project, Mediator method and XML technology, the paper discusses the problem of the multi-source heterogeneous data integration, which intend to resolve the heterogeneous among data sources, semantic conflict and global access performance etc. The goal is to provide users a global transparent query function. Therefore, the paper focus on the global data modeling, registration of local data source, pattern conversion, query decomposition and transformation, combination of query result and so on.The paper implements a data integration system in the context of a certain project. Preliminary experiments show that this systems complete data integration for heterogeneous data source including semi-structured and structured data sources. The work satisfies the requirements of the project and can guide the actual engineering to a certain extent. |
PDF Full Text Request