XML Schema Based XML Index

With the rapid development of Internet and e-commerce, especially the wide useof Web service, XML (Extensible Markup Language) is used as data representationand data exchange standards. Because of its flexibility and cross-platform, XML canbe directly used for the Internet and supporting a variety of applications. Currently,XML type data has become the main form of data representation. Therefore, XMLdata management technologies, especially XML data query technology has become aresearch hotspot, which gives rise to new requirements and challenges for XML queryprocessing technology.An XML document can be modeled as a tree. XQuery, a generic XML querylanguage, whose query expression, XPath, can also be modeled as a tree which can benamed as tree pattern. Thus it has become one of the key technologies in XQueryimplementation which make use of pattern matching algorithm efficiently to extractthe results meeting the requirements of tree pattern from the XML document. Butwith the dramatical increasing in the size of XML data, navigational query can notmeet the requirement of the high performance for XML query processing. Due to thisrequirement, many non-navigational methods about tree pattern matching have beenput forward in XML database research field, in whch one of them is performing queryon XML document with index. The purpose is to make a pretreatment for XMLdocuments, and then perform the query on XML document by making use of efficientpattern matching algorithm. Because of different types of index, the ability of thequery may also be different. Such as the granularity and structure of index will have abig influence in the efficiency of query processing.In this paper, we propose a new XML index which is created based on thecombination of descendants’ labels. The granularity and structure of index are takeninto consideration. According to the structure of tree pattern, we only process parts ofthe XML document which have relationship with tree pattern and thereby, it willreduce many operations making no contributions to query. We also propose a newquery processing algorithm which can make full use of the new index structure. In thequery, by combing the structure of tree pattern with the index, we only need to selectparts of document which meet the requirement of tree pattern. In this way, we canreduce the intermediate size generated in the processing of query and improve thequery efficiency.Experitments show that the index has a reasonable size and easy access property.With different type of data sets, we make a test about time and space for this indexbased query processing algorithm. The result demonstrates a good utilization on memory which can reduce a large number of operations making no contributions toquery and speed up the procedure of query processing.