Pushdown Transducer Based Query Method Over XML Streams

In the last ten years, the Internet technology develops quickly. The number of Internet users keep increasing. And network applications become more and more universal. The web data vloume is also increasing quickly. As the factual standard of data exchanging and sharing, XML(extensible markup language) has been widely used in various Internet platforms. XML has become a popular data storage format. Lots of areas store information in XML format. In many cases, the size of one XML document is already far beyond the limited computer memory capacity. How to get the information we need from huge amounts of XML data is a problem to be solved.Traditionally, we load the entire XML document into the memory, then get the query result directly from the memory. Obviously, this method is not suiTablefor the data that is much more larger than the memory capacity. Therefore, another form of XML data- streaming XML data gradually gets our attention. At the same time, the characteristics of streaming XML data put forward new challenges to query and processe it. Characteristics of streaming XML data are as follows: tag/token, as the unit of the data, arrives in real-time; the processing system is not able to control the arrival sequence of the data. The data may continully arrive and will not stop. The XML stream data processing system generally uses XML parser parse the data, and uses XPath or XQuery to describe user’s query demand. To reduce the memory usage, one data will be discarded as soon as it is processed in the system.At present, the problem of querying streaming XML data has become a hot issue in the streaming data research field. The goal of the research is find the information users need effectively in the limited memory. There has already been lots of methods for solving this problem. However, these methods still can’t satisfy users’ demands. They are either not able to support the complex query or not effective enough. To make up the shortcomings of the existing methods, this dissertation proposes a new XML streaming data process method. It is a pushdown transducer based method for XPath language. It supports the core part of XPath, including PC(parent-chidren) and AD(ancestor-descendant) axes and multiple predicates with nesting step(i.e., predicates which are either value based predicates or node tests). This method uses XPath step as the design unit. Therefore, the state number will not soared. It can find all the results only by traversal the data once. It is able to handle complex queries with high performance.This is a high-efficiency method and can meet the demand of XML streaming data processing better. It makes up for the deficiency of previous pmethods.