Research On Twig Pattern Query Based On XML Data

With the rapid development of information technology, Extensible MarkupLanguage(XML) has been a de facto standard of network data storage and exchange andmore and more xml data become the carries of information sharing. The emergence oflarger amounts of XML data promotes the rapid development of XML database. As thecore component of XML query processing, XML twig pattern query has been a hot spot inthe research community. This paper studies the efficient twig pattern query, and the mainresearch contents are as follows:Firstly, through the analysis and comparison of the existing twig pattern matchingalgorithm, we summarizes the field of each algorithm and the problems. TwigStackalgorithms and TJFast algorithms can achieve optimality in case of twig query with onlyancestor-descendant relatioinships between branching nodes and their children, whenprocessing the query, we get the partial results and the merger of them, which may bring alarge number of middle results and bring additional cost. In addition, bescase of that wecan not get the information of element directly through the coding, the cost of processing asingle element is relatively high which affects the overall efficiency of query.Secondly, the paper proposes a new coding scheme by adding path index in Deweycoding and we can get the label of element node and ancestor node whitout dependentingthe Finite state transducer that reduces the the cost of processing element through thecoding. For the problem of low efficiency of the existing methods, we propose a new twigpattern query algorithm called TJPath that reduce the the cost of processing a singleelement.Finally, we propose a experimental system and implement TwigStack, TJFast andTJPath algorithm. Through the comparison of the rich experimental results on the numberof scan elements, processing time and scalability verifies the effectiveness of the...