Research And Implementation Of The Full-text Retrieval System In XML Databases

With the wide use of XML, there is a large number of XML data. In order to manage these XML data, The XML database was generated. CoSQLRX is a relation-XML daul engine database management system, which can deal with both relation query and XML query. This thesis focuses on the study of full-text retrieval on XML. By analysing kinds of full-text retrieval on XML, this thesis proposes a full-text retrieval system on XML layout in CoSQLRX system.Firstly, some related work of full-text retrieval is introduced. The syntax of full-text retrieval on XML, and the relation with XQuery, and the CoSQLRX database are introduced in this thesis.Secondly, this thesis gives a detailed introduction about our design of full-text retrieval module, including the design of the parser and all the operators of the plan. We also design the algorithm of all the operators.Thirdly, this thesis proposes the logic query plan optimization strategies and physical query plan optimization strategies for typical queries after analyzing the query. We also introduce and analyze the XML labeling relation with the scoring in the full-text retrieval, and we analyze the labeling problem on updating XML. The DCU, which is a new menthod of labeling XML, is proposed in this thesis. The menthod that how to design each module is implemented in the CoSQLRX database, consist of the expanding the original syntax rules, the registration of the new operator for the full-text retrieval. This thesis introduces the scheduling of each operator function relationship.At last, after analyzing the experiment results, the design of the full-text retrieval of this thesis can be well implemented. The experiment result shows that our optimization strategy can improve query efficiency.