【作者】 姚文隽；

【导师】 金远平；

【作者基本信息】 东南大学 ， 计算机应用， 2004， 硕士

【摘要】 XML近来已成为Web上表示和交换数据的统一标准。随着XML在信息集成、电子商务等方面的广泛应用,大量的数据信息需要在关系模式和XML模式间转换:从外部获得的XML数据需要转换为关系数据存储在数据库中;而内部的关系数据又需要映射为XML数据在Web上传输和交换。因此,基于关系数据库的XML存储技术和XML出版技术成为数据库领域中的新兴研究热点。本文对现有的存储技术、出版技术作了全面比较,并就这两类技术中的共同难题——将XML查询转换成等价的SQL查询的方法进行了深入研究。我们综合比较了XML存储技术和XML出版技术中多种查询转换方法,得到结论:在XML存储技术中,查询转换仍然是一项重要的挑战,目前尚无很完善的转换方法。且由于这种转换依赖于关系模式生成技术,各种生成技术都必须构造自己的查询处理器。而XML出版技术中,已存在较完善的查询转换方法,且有的方法具有一定的通用性。利用这一点,我们采用将出版技术领域中较为成熟的转换方法应用于存储技术领域中的思路,具体设计了一套完整的方案。我们通过构造SilkRoute中的关键结构View Forest,将SilkRoute中查询转换算法应用于多种XML存储技术之中,作为XML存储领域的通用查询处理器。但SilkRoute的算法无法处理XML中的顺序信息。针对这一弱点,我们又改进了SilkRoute中的数据结构和算法,使之能正确转化基于顺序信息的XQuery查询。最后,本文综合评述了该方案的优点和存在的缺陷,提出了需进一步开展的研究工作。更多还原

【Abstract】 XML is emerging as a standard format for both data representation and data exchange in the web. Various kinds of applications using the XML format have been developed, such as information integration, e-business, etc. Thus, huge data has to be transformed between relational format and XML format, i.e. the data obtained from web in XML format need to be stored as relational data in RDB, while the data in relational database need to be published as XML data for transmission and exchange in the web. So more and more research on using RDB to store XML data and publishing relational data in XML have been done. On the basis of comprehensive comparisons among existing technologies in XML storage domain as well as XML publishing domain, this thesis focuses on in-depth investigation of a challenging problem in both domains, i.e., how to convert XML queries over the XML documents into equivalent SQL queries over the tables.Having studied a great deal approaches to this problem in XML storage domain and XML publishing domain, we draw a conclusion that in XML storage domain, it is still a challenge to convert XQuery efficiently and no existing technologies to our best knowledge can solve this problem soundly. Moreover, this kind of conversion relies on relational schema generation,so each technique should build their own query processor. However, in XML publishing domain, there are many methods that can effectively translate most XQuery expression into an equivalent set of one or more SQL queries, and some of them are quite adaptive. Consequently, we propose a generic approach to this problem of XQuery processing in XML storage domain by means of using conversion methods from the XML publishing domain in the XML storage domain.We make it possible for most XML storage technique to use the algorithm of transforming XQuery in SilkRoute as its own query processor by constructing view forest which is the key data structure in SilkRoute. Moreover, SilkRoute has its own limitation. It can’t deal with any XQuery expression based on the order of XML data. So in our approach, data structures and algorithms in SilkRoute are also ameliorated to be capable of transforming XQuery statements with sequence information. Finally, we give a conclusion of the advantages as well as drawbacks of our approach, which will guide our research work in the future.更多还原