کد مقاله | کد نشریه | سال انتشار | مقاله انگلیسی | نسخه تمام متن |
---|---|---|---|---|
394804 | 665844 | 2012 | 14 صفحه PDF | دانلود رایگان |
More and more XML data is generated and used for data exchange. In this paper, we address the problem of filtering XML documents with large number of XPath expressions, which may contain ‘ancestor’ and ‘parent’ axes. XPath expressions with these axes are more powerful and flexible for users to describe their interests in publish/subscribe systems. First, we analyze the characteristics of the ‘parent’ axis and propose a series of rules to eliminate it in XPath expressions. Then we propose a new index structure called NIndex, which is designed to efficiently store and index large number of XPath expressions. NIndex offers several features which make it especially attractive for the large scale selective dissemination of information, including the ability to handle complex XPath expressions with ‘ancestor’ and ‘parent’ axes, and efficient pruning. Based on NIndex, we design a new filtering algorithm with low complexity for our problem. Our experiment results show that our algorithm performs well across a range of XPath expressions and documents.
Journal: Information Sciences - Volume 210, 25 November 2012, Pages 41–54