Research On Keyword Query Over Relational Databases

With the rapid development of database and information retrieval, how to achieve information retrieval functions in relational database management system by which the normal user or web user can access relational database effectively only through the simple querying keywords without understanding the relational database schema and the write way of SQL query has been the hot topic in many researchers recently. Fully assimilated their research results, there is a focus on the query efficiency and query results of querying keywords. This paper chooses several problems of the existing method, conducts intensive research, and presents the solution method.First of all, the basic knowledge of keywords querying over relational databases is introduced. The idea of the keywords querying method based on seaching database schema graph and data graph system over relational databases is stated, then their advantages and disadvantages are analyzed.Secondly, aimed at the problem that the query efficiency is low and the query results can't be satisfied in the relational database keywords querying system, a new structure which is facing the top-k hierarchy database keywords query system architecture is put forward, and the architecture is instantiated by using the IR-style ranking expressions,candidate network pretreatment method based on the index,bound real-time filtering threshold top-k search algorithm which are proposed in the paper. And the keywords query prototype, Prober, is achieved. Then Prober is expounded in depth from data model,query language,architecture,query processing,IR-style ranking expressions,query algorithm and so on.Thirdly, the relational database keywords query technology mainly utilize the syntax matching the keywords, while not utilize the semantic relation among data to match, whose results can't satisfy the public. Then aimed at the problem, and combining with the concept of Ontology, a semantic expansion method of keywords query based on Ontology over relational database is proposed, which extends the query keywords provided by users to the semantic keywords based on Ontology.Finally, the experiments validate the above-mentioned methods of keywords query, verify the effectiveness of these methods, feasibility and superiority.