Research And Implementation On XML Index In Relation-XML Dual Engine Database Management System CoSQLRX

This thesis describes the research and implementation of the XML indexes which are based on the relation-XML daul engine database management system CoSQLRX. The daul engine database management system is the database management system which can deal with relation data and XML data concurrently. With the developing of the XML related technologies continuously, application fields of XML data becomes more and more widely. Many fields have the application requirements of the unified management of relation data and XML data, but at present there isn’t a influential relation-XML dual engine database management system inland. So the implementation of the CoSQLRX has important realistic significance. Based on the relation database management system, CoSQLRX adds in the native XML storage and query processing technologies over XML data. It could handle the data management of many fields, such as the enterprise informatization and the electronic business.Creating XML indexes is an important means to accelerate XML query. In CoSQLRX, XML data index is a very important module. This thesis mainly proposes the design and implementation of this module.First, this thesis proposes the conception of XML data index. It also introduces its background and current status. This paper analyzes several kinds of XML data indexes and their features.Second, this paper proposes the design of the XML index module. It explains how the XML data is stored in the CoSQLRX system, the design of the structure of XML index, and the processes of creating, updating and scanning a XML index. In the CoSQLRX system, the structure of a XML index mainly contains two parts, which are XmlInfoRelation and XNBTREE. When the system creates a XML index, it reads data from the native XML storage and collects useful information, such as the path information, the value information, etc. Then the system stores these informations into the XmlInfoRelation and XNBTREE. When the system executes a XML query processing, it analyzes the query statement, and then call the suitable XML index. The system searches the nodes which are accord with the query conditions. If it finds one, it will take out the corresponding informations stored in the XmlInfoRelation and make further analysis to design whether this node is the correct result of the query statement. Based on the thoughts above, in this paper we design and implement four kinds of XML indexes, which are path index, value index, ancestor index and twig index.At last, this thesis presents the implement of XML index module, and uses the XMark test bench to do some tests on the creating and querying of these XML indexes. The experiment results shows that the XML queries are effectively accelerated after creating XML indexes.