The Research On Database Schema Matching System

With the development of information age, the demand of discovering semantic consistency between different schemas has been imminence increasingly. Schema matching as the first step of schema operation, it has a crucial effect in many fields, such as Data Integration, Data Translation and Model Management. This paper analyzes the current situation of the domestic and international schema matching problem, and researches for the problem of complex schema matching from a completely new perspective.At first, the two kinds of existing complex schema matching methods are introduced. This paper points out the merits and demerits in each method through analyzing the function of each process module and the interrelated technologies in the two methods in an all-round way.Secondly, on the basis of the above research and existent solutions, the extended complex schema matching system CSM is proposed. It filters some unreasonable matches on data types and values by preprocess and clustering process, and employs a set of special-purpose search procedures in match generator to explore a specialized portion of the search space and discovers 1:1 and complex matches; It estimates candidate matches and selects optimal candidate matches by using similarity estimator and match selector respectively; According to the problem that there are opaque columns in the schemas being matched, it can apply complementary matcher to find matching relations between opaque columns further more. Thereby it can discover more general, reasonable matching pairs.Moreover, depending on the idea of using a corpus to accomplish schema matching, a new schema matching method named corpus-based complex schema matching is proposed. According to the problem of the lack ofsufficient evidence in the schemas being matched, it can use a known corpus of schemas and mappings between schemas to augment the evidence about the schemas being matched, thereby improves the matching recall and precision.Finally, we analyze the matching process of the two schema matching methods as above using instances, and prove their feasibility and validity by theory and experiment.