Toward a Streaming Evaluation Model for XPath

• Main Authors: Li-zhen Liu, 劉麗貞
• Other Authors: Hahn-ming Lee
• Format: Others
• Language: en_US
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/57317196935632103424

碩士 === 國立臺灣科技大學 === 資訊工程系 === 96 === Extensible Markup Language (XML), a markup language proposed by World Wide Web Consortium (W3C) in 1996, has become a standard format for the exchange of information among various applications on the Internet. With the popularity of XML, the amount of XML document is growing rapidly, when the document which is multiple times larger than available main memory, it will be failed to use DOM based approach. We present a streaming evaluation model for evaluating a XPath expression (involving forward axis, such as child, descendant, following, following-sibling and backward axes, such as parent, ancestor, preceding, and preceding-sibling axes) on a XML document. Besides, we propose an approach which can convert a XPath expression into a corresponding abstract syntax tree suited for evaluating on streaming XML document. Of course, not all XPath expressions can be evaluated this way. We show how to analyze a XPath expression to tell if and how it can be streaming evaluated, and if cannot, how to break the evaluation into multiple passes where each is in streaming manner. Furthermore, we view this abstract syntax tree as computation module against very large XML document. In the most practical case, the processing time is linear to the XML document size and memory is linear to the depth of XML document excluding backward axes. Experiments showed that proposed approach with a good efficiency on execution time within as little memory.