The Research And Improvement Of A Native XML Database-Xindice

With the rapid development of the World Wide Web, the application scope ofXML is expanding, as well as XML databases are becoming the focus of public.However, Relational Database Management System (RDBMS) couldn't support XMLvery well. Because it transforms XML to table, while the table transformation bringsperformance issue.Native XML database is the current hot issue in the research field of XMLdatabases. And in Native XML database, query and storage strategy is very important.In this paper, writer studies an open XML database—Xindice, analyze its multi-layerstorage model as well as its page storage strategy and B tree index structures. On thisbasis, the deficiencies of the Xindice storage strategy and the query strategy arepresented. Against its deficiencies, the paper made the following further research.In the storage strategy, when adding a new XML document in a collection,Xindice database allocates "idle" pages to store data files. If there are no spare pages,create new ones; however, when deleting XML documents from the system, thedatabase only marks "idle" in the pages to be deleted, without really releasing itsoccupied space. In this paper, based on experiments, In order to resolve this problem,the paper designs two new strategies, tail page truncating strategy and page movingstrategy, and implements them, which help Xindice release the disk space held by thefree page and increase the utilization rate of the disk resources. Also, in order toprevent moving pages quite frequently in page moving strategy, the paper bringsforward tail page moving strategy to reduce time cost in releazing free page.In the query strategy, the query language of Xindice is XPath, not supportingXQuery language. XPath is quite limited in query, which is not able to support group,sort, connect functions and can not query connected tables. This deficiency bringsdown the flexibility in Xindice query. This paper designed XQuery expressiongrammar, use JJTree and JavaCC tool for XQuery expression grammar to generatelexical syntax analyzer, which is used to identify imported XQuery expression syntaxstructure, and generate the corresponding syntax tree. According to the syntax tree, the clauses of the query XQuery expression were dealt respectively and then the ultimateXML query result was generated. The efficiency of XQuery expression in this paperhas been proved in experiments.