The Query Of XML Data Based On IPE Encoding

With the rapid development of the Internet and the increase of Web information,traditional HTML is not satisfied with the need of information exchange.Because XML is unconcerned with platform,well-format and easy to expand,which has good interactive nature and strong semantics,etc.XML has become the universal format for the representation,exchange and storage of information over the Internet.The storage and query of XML become more important.Relational Database gradually matures at present.Combining XML with Relational Database is a feasible and effective method,and it will be a hot spot.To effectively support structure query,various encoding schemes of XML data have been brought forward at present.To decrease visiting database and connecting operation,the index technology which can improve query efficiency is introduced.About structure query,this paper encodes nodes in the XML document tree,and the calculation of XML structure query is switched to structural join through codes.Firstly,this paper does a research on the storage of XML data based on Relational Database and the querying process of the path expression, so IPE encoding is proposed.IPE encoding is an improved XML prefix encoding.The prefix encoding preserves the element path information,and the node's parent code is the prefix of the node's code.The sign"."separates the layer of node's code,and it takes increasing the code length and memory space as price.So Improved Prefix Encoding(IPE) is suggested,it cancels separated sign,adopts corresponding table,solves the confusion of seperating code's layer and reduces code length when the XML document tree has more children or more siblings.IPE encoding supports XPath query and it need the comparison of code, the judgement of relation such as ancestor/descendant,preceding/following,etc in the query.The discriminance method of XPath query axis based on IPE encoding is provided.Secondly,inverted index tables including Pc,Ad and Name are built,which are recorded nodes's parent IPE encoding,ancestor IPE encoding and IPE encoding with identical element name(or value)/attribute name(or value) respectively. During dealing with path query adopted IPE encoding,inclusion relation algorithm,holding relation algorithm and branch path query algorithm are brought forward,they effectively improve query speed.Finally,the validity of algorithms in this paper is indicated.The experiment result proves that the algorithms in this paper are better than Stack-Tree-Desc algorithm.