Research On Consistency And Index Based On B Tree Of Temporal XML

With the rapid development of network applications, XML has been widely applied to Internet IR systems, digital libraries, data integration, Web Service and other fields, which makes XML become the primary data format. Therefore, how to extract useful information from XML data has become a hot research area.Time is a ubiquitous and objective attribute in the nature, and all information has its corresponding temporal attributes. With the development of database and information technology, it is more and more urgent to process temporal information efficiently. Temporal XML, as an emerging research branch of XML technology, is proposed to meet to the requirement of processing such temporal information, which plays an important part in e-commerce, e-government and other fields. Temporal XML data model could express temporal information more intuitively, which is a significant advantage compared to temporal relational database. Most of the researches nowadays are focused on the mechanism of data modeling and querying, and the expression of temporal information with XML, rather than physical storage and indexes. However, traditional XML indexes don't do a good job in improving the query efficiency of temporal XML documents. Consequently, it is essential to introduce some special indexing technology.In this thesis, we study the process of temporal XML modeling. Up on this model we propose an indexing scheme based on B tree and optimize queries relevant to time. Firstly, we introduce the temporal XML model based on XPath and analyze the inconsistencies which might appear in the process of modeling. Based on our analysis, we propose algorithms for checking and fixing inconsistent documents. Secondly, we propose a new temporal XML index model-TBtree-Index, which includes temporal XML coding and temporal interval TBtree index; We give the descriptions about the approaches of ORDPATH coding scheme and updating for temporal XML, and the algorithms for building temporal interval index based on B-tree in detail. Finally, we discuss the applications of TBtree-Index in queries. Our experiment results show that TBtree-Index is feasible and has its advantages in efficiency compared to other indexes.