XQuery Processing Based On Relational DataBase

XQuery processing based on RDBMS (Relational DataBase Management System) is one of the most important parts in the management on XML data with RDBMS. In order to develop XQuery processor for DaMeng RDBMS, Some researches have been done on the following points, query's transformation from XML to SQL, interpretation of XQuery script, optimization strategy on XQuery and so on.On the base of XPath query expression, XQuery language adds many other expressions, such as FLWR expression, to do complex query or reconstruction on XML data. According as this characteristic of XQuery language, the XQuery processing is divided into two logic-independent parts: the first one is XPath query's processing, the second one is the interpretation of XQuery which is based on XPath query. In the second part, do some optimization strategy to accelerate the query.There are three key techniques in DaMeng's XQuery processing. The first one is the transformation arithmetic which transforms XPath query into SQL query, this arithmetic makes use of the XML document's structure information to accelerate XPath query. The second one is interpretation technique for XQuery script. In the syntax phase, a recursion-down way is used for XQuery script's syntax analysis. In the semantic phase, there are two kinds of semantic analysis way have been implemented successively for interpreting query. One is based on recursion-down while the other is based on the Query structure. The latter is an improved way based on the former. In these interpreting processing, an engine based on simulative- circulate is designed for FLWR query which is used most in XQuery. The last one is the strategies that optimize XQuery processing. There are two strategies have been carried out for XQuery processing optimization: one is the relation's processing between XPath expressions, the other is the rewriting of simple FLWR query.In the experiment, a benchmark system named XMark was used for testing XQuery processing's performance. The result indicates the key techniques listed in the above paragraph have achieved the expectant purpose and the XQuery processor's performance is up to the system's purpose.