Study And Implementation Of Data Integration Technology In "One-Stop Query System"

With the rapid development and extensive application of information technology, a large scale of distributed and heterogeneous information sources emerges. These various types of sources are of different ways of expression and storage as well as belong to different hosts, so that they are not fit to be shared, which brings great difficulties to the comprehensive utilization of information. Information integration is the important way to effectively utilize these forms of information, and the data integration service is the foundation of this process. The traditional ways of data integration have solved the problems of heterogeneous system, structural isomerism and grammar heterogeneous to some extent,but they seemed incompetent to solve the problems of data consistency maintenance, model effectiveness detection and data quality management.After more than ten years of information construction in the tax administration field, there are a lot of information systems needed to be integrated for applications. The thesis designs out data integration model based on the data warehouse and the middleware according to the practical needs of the tax administration field, and furthermore makes an exploration over the key technology with these basis trials, finally accomplished the"One-Stop Query System". The main content of the study is as follows:Firstly, the thesis analyzes the characteristics of the data application and the practical needs of the data integration in the tax administration field, and designs out the data integration model. Also, the model respectively realized the functions of data consistency maintenance, the detection of the mode effectiveness and the quality management of the global data.In the study of data consistency maintenance between source database and global database, this thesis firstly analyzed the research status quo of the data consistency maintenance strategies, especially the data consistency maintenance strategies based on database journal. The thesis, with the support of sources database to trigger, suggested some strategies to maintain the data consistency based on user journal considering the defects such as low efficiency and complexity of the present models.In the research process of the semantic mode collision detection caused by source database updating, this thesis, based on data dictionary model, proposes a method of comparison to locate failure model, thus completes model validity testing and failure mode accurate positioning.In the applications of data integration system, global data quality plays the crucial role in the success of the whole system, and this thesis designed the data quality management module for this purpose. Through data quality monitoring in the global database, erroneous data are detected, discriminated and repaired, and then data quality"zero error"management is realized.Based on the above study, we fulfilled the implementation of the"One-Stop Query System"and verified the feasibility and effectiveness of the model. The successful trial operation of the system in our province'tax administration field has laid a solid foundation for online operations of the system in the vast scale.