Research On Schema And Optimization Of Heterogeneous Data Resources Aggregation

The resources of enterprise information are distributed and heterogeneous, now we can access the data of data source with the unified access interface as the general data access provided unified method to access multisource and heterogeneous data source, then mapping the source to an target schema to build a visual flow model,last generating the result which is desirable for the users.It's complicated, time-consuming and fallible for engineers to define schema mapping with visual modeling tools to finish heterogeneous data aggregation. Facing the challenge, this paper depends on "XBus Data Aggregation Platform" and does some research on some issues, which contain the schema of the data aggregation and the problem of optimize the query. The main contributions are as followed:Firstly, it is proposed the method to build visual flow model based on Schema Mapping. It presents the conception of the mapping operation based on Schema Mapping, brings the definition and implementation of a set of mapping operation to extend the semantics of the mapping schema. It can automatically maintain the semantics between the source and the target through the mapping operation,users don't need to know the semantics of each operation definitely, so it can reduce errors during the aggregation. Finally it presents a method to build the data aggregation, and supplies a data model and a algorithm of data flow modeling.Secondly, it is proposed the optimization method of data aggregation.This paper does some analysis on the efficiency and capability on the data aggregation based on XQuery, and presents an optimization algorithm for XQuery based on projection and the conception of projection path and projection document,then gives the algorithm of path analysis and loading algorithm which can increase the efficiency of the XQuery engine which is restricted to the main-memory.Finally, the data aggregation platform based on the researches above is designed and implemented. We implement a data aggregation (XBus) with result of the researches above, creat a visual flow building tools based on mapping operation,and opitimize the XQuery parser,last make some experiment to compare it with the related product and illustrate the usability and high efficiency of the platform.