异构数据源集成系统中查询重写的研究

发布时间：2017-12-28 04:27

本文关键词：异构数据源集成系统中查询重写的研究　出处：《哈尔滨商业大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着计算机技术的飞速发展与广泛应用,数据量已经无法用"多"来形容,由于各行各业对数据的需求不同导致数据的存储方式,数据结构等多方面都存在差异,因此形成了大量的异构数据源。但对于用户来说这并不是他们想要的,用户通常希望通过提交一次查询就可以得到需要的数据,异构数据源集成系统应运而生。其中,查询重写技术在异构数据源集成系统中扮演着极其重要的角色,集成系统正是通过查询重写技术将用户给定的基于全局模式提出的查询语句进行重写,以实现从异构数据源中获取结果并反馈给用户。查询重写技术与数据集成、查询优化等问题都密切相关。课题针对异构数据源集成系统中的查询重写问题做了以下研究。首先,课题对三种经典的查询重写算法,Bucket算法,Inverse-Rules算法,以及MiniCon算法进行了深入的研究,并分别指出了以上三种算法的不足之处。着重探讨研究了 MiniCon算法,并在该算法的基础上提出了一种改进算法,即基于路径优化的MiniCon算法。该算法在传统的MiniCon算法的基础上,增加了一步路径的优化,通过比较查询视图中相关字段的数据有效比例,对查询路径进行优化,以达到提高查询效率的目的。其次,课题介绍了三种传统数据集成方案,即联邦数据库法,中间件法以及数据仓库法。并以中间件体系结构为基础,融合JSON技术,设计了一种异构数据源集成框架,采用稳定的三层结构,包括展示层,中间层及数据源层,其中,中间层为此系统的核心,查询生成、查询重写等都在中介层中实现。最后,课题将传统的MiniCon算法与改进的基于路径优化的MiniCon算法应用到上述设计的异构数据源集成系统中,并采用河南世纪联华超市的数据,对两种算法的查询速率进行了比较以证明改进算法的正确性与优越性。
[Abstract]:With the rapid development and wide application of computer technology, data volume can not be described by "multi". Due to the different needs of data in different industries, there are many differences in data storage mode and data structure. Therefore, a large number of heterogeneous data sources are formed. But for users, this is not what they want. Users usually hope to get the required data by submitting a query, and heterogeneous data source integration system arises at the historic moment. Among them, the query rewriting technology plays a very important role in the integration of heterogeneous data sources in the system, the integrated system is through the query rewriting technology will be given by the user query based on global schema rewriting, in order to achieve from heterogeneous data sources to obtain results and feedback to the user. Query rewriting technology is closely related to data integration, query optimization and so on. The following research has been done on query rewriting in heterogeneous data source integration systems. First of all, three classical query rewriting algorithms, Bucket algorithm, Inverse-Rules algorithm and MiniCon algorithm are deeply studied, and the shortcomings of the above three algorithms are pointed out respectively. This paper focuses on the study of the MiniCon algorithm and proposes an improved algorithm based on the algorithm, that is, the MiniCon algorithm based on the path optimization. Based on the traditional MiniCon algorithm, the algorithm adds a one-step path optimization. By comparing the effective proportion of related fields in query view, we optimize the query path, so as to improve the efficiency of query. Secondly, we introduce three traditional data integration schemes, namely, federal database method, middleware method and data warehouse method. And the middleware architecture based on integration of JSON technology, design a framework of heterogeneous data source integration, adopt three layers of structure stability, including the presentation layer, middle layer and data source layer, the middle layer to the core of the system, are implemented in the intermediate layer, query rewriting query generation. Finally, issues of the traditional MiniCon algorithm and improved MiniCon algorithm is applied to path optimization based on the design of heterogeneous data source integration system, and the Henan Century Lianhua supermarket data, the two algorithms of query rates were compared to prove the correctness and superiority of the improved algorithm.
【学位授予单位】：哈尔滨商业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.13

【参考文献】