Contains The Predicate Of The Xml Path Expression Query Processing Methods

The characters of flexibility and expression ability make XML be the facto standard of information exchange on the Internet. Many companies and research institutes dedicate their time and resource into XML research. Twig pattern match is one of the most popular and fruitful direction. Many algorithms and solutions are given to tackle the problem of path or twig match, Such as structural join. Compare with that, the keywords search on XML database or documents are less focused. Especially the keywords search with path requirements. As far as we know, most of keywords search algorithms focus on the LCA problem. This dissertation uses the path match algorithm as the framework with additional index to fulfill keyword search.This dissertation gives two algorithms. The first one is BFX, which combines the length of XML DATA node with node ID as the index. This method accumulates length information in the process of structural match to locate the contents of node for the efficient evaluation. More than that, for enhancing its adaptability, we take partition strategy to repository. the improved algorithm separates the index into two level, and store the content in to different barrel based on the code range, which releases the strict sequential requirement of the repository of node content.The second algorithm is IFX, which uses the keyword invert index to link the nodes and the keywords. In the query process, the answer nodes from path matching will be sort-merged with the merge result of the keywords inverted list. The improved IFX index adds the range information of the node into the index, enabling the keywords merge result infect the path pattern matching process. This character diminishes the pattern matching cost by reducing the stack operation.