A PARSING CACHE FOR HIGHLY EFFICIENT XML DATA PROCESSING

摘要

XML (extensible Markup Language) documents are relatively large. Therefore processing such documents can be very costly. Meanwhile traditional filesystem buffering and caching mechanisms are not designed particularly to support XML parsing. Hence, redundant parsing problem causes the current XML parsers and query processors not to be efficient when processing large documents. Also, due to the data structure and algorithms used in the current XML tools, parsing is not fast enough or even providing required resource (i.e. memory) may not be feasible. Furthermore, XML tools are not designed to provide a sharable parse tree for multiple users, and users have to limit themselves to specific tools to access the parse tree. To overcome these limitations, we introduce a "Parsing Cache", which uses a well-structured parse tree, an efficient parsing algorithm and a user-cache interface. It provides a resident parsing storage specially for XML documents. Due to small memory usage, parse trees can be resident for further processing. The user-cache interface enables multiple users to share the parse tree. The parsing and query time are significantly improved compared with the current XML parsers and query managers.