Research And Implementation Of XML Storage Technology In Relational Databases

XML(eXtensible Markup Language) is an important Internet technology following the HTML.It has become the new standard for Internet data representation and exchange.The volume of XML data is increasing with the popularization of Web application.It is critical to have efficient mechanisms to store XML documents to exploit the full power of the new technology. Depending on the powerful and reliable data management services, using TDBMS to store XML documents becomes a considerable approach to solve the problem. However, due to the mismatch between the XML's tree structure and flat relational tables, mapping XML document to relational database is an involved problem. So it becomes a pop research direction.This paper studies a number of strategies for storing XML documents in RDBMS and illustrates their advantages and disadvantages, adaptability and performance. Then an XML-to-relational prototype system X-R was proposed that is comprehensive and generic.Mapping in X-R are defined through annotations in an input XML Schema.Relational schemas are generated through parsing the annotated XML Schema, the mapping information is held in the mapping repository. The document shredder accepts as input a document and accesses the information in the mapping repository to generate the records and populate the tables in the relational schema. The mapping repository is also accessed by the query-translate module, which generate SQL queries from XML queries. Modules are independent.Various mapping strategies were taken into account of designing annotations. By using various annotations the system are capable of using a wide range of existing mapping strategies and easily extensible to new mapping strategies. Considering of various applications the annotations in X-R can express the strategies of edge table, basic inlining, shared inlining, hybrid inling techniques and so on which can satisfy various requirement. X-R is portable and can be used with any relational database backnd.The system is generic,fiexible, easy to use, portable and extendable.In the end of this paper, a complete example is given to illustrate and test X-R and the result is satisfying.