The Model Construction And Optimization Of XML Document Filter System

Extensible Markup Language (eXtensible Markup Language, XML) has come being not long ago, however, it has been the standard for data description and message exchange, and widely applied in the many emerging systems. At present, some forms of XML are adopted in the Internet, the management systems in enterprises and government departments, personal computer to store data and exchange information. Therefore, XML data query and filter becomes increasingly important. XPath expressions are usually used in XML documents filter system to represent personalized information of users. As automaton can modeling perfectly XPath. Therefore, the current XML filter system is generally based on XPath and automaton.The traditional XML filter system needs to scan and parse every element in an XML document, so a large proportion of the filter systems are parsing the XML document. This increases filtration cost and affects the efficiency of filtering. For this reason, this paper puts forward a new XML index mechanism, called stream index, to skip worthless elements and their sub-elements. Therefore, the elements which match elements'position can be quickly located, and to the largest extent, eliminating elements that are not guaranteed to be part of any match is able to avoid an amount of redundant processing. Moreover, the proposed method is extensible, and can improve the treatment efficiency indeed. Preliminary experiments show all algorithms have linear time and space complexity, and are capable of improving the treatment efficiency of XML filter.The above technology is about processing XML documents, but XPath expressions, especially which contain predicate and nested path, are also the fundamental research problem of XML filter system. This paper proposes a new modeling idea to handle this situation. It is that each node of XPath expression is modeled into a trigger condition between automata states. This idea uses a single automaton, so it need not define an automata or filter for each branch structure. We also optimize the model which contains boolean operator "AND", and present the application of the thought in XML query and filter. The experiments show that our new method is better than existing methods in processing predicate and nested path.