Research On Algorithms For Rewriting Queries In Data Integration

Query rewriting is a basic problem in database query disposal procession. It is closely related to the query optimization, maintenance of physical data independence, data integration, semantic data caching, data warehousing and decision support. With the development of data integration, query rewriting problem in data integration become research hotspot. This paper analyzes the current situation of the domestic and international query rewriting problem in data integration, and researches for the problem of query rewriting from a completely new perspective.At first, the elementary theory of data integration and basic conception of datalog language are introduced. Through analyzing two query rewriting algorithms, this chapter points out the deficiencies of two algorithms.Secondly, through analyzing traditional MiniCon algorithm, this chapter points out the deficiencies of the algorithm. Constraint-based MiniCon algorithm is proposed. This algorithm adds a step of view selection. As a result, the problems of missing query rewritings and generating redundant query rewritings in traditional MiniCon algorithm in presence of domain constraints are solved, and the accuracy and completeness of the algorithm is ensured. This chapter also gives the corresponding algorithmic analysis and the example analysis.Thirdly, through analyzing aggregate query rewriting algorithm in exitence, this chapter points out the deficiencies of algorithm. The MiniCon algorithm thought is integrated in the aggregate query rewriting. Count_Rewriting algorithm and Sum_Rewriting algorithm are proposed. The accurate proof and the example analysis are carried on to the Count_Rewriting algorithm.Finally, the experiments are carried on to Count_Rewriting algorithm and exhaustion algorithm. Through analyzing experiment results, it proves that the algorithm efficiency in Count_Rewriting algorithm is higher than the algorithm efficiency in exhaustion algorithm.