Supporting efficient query processing on compressed XML files

XML has been widely accepted as the de facto format for data representation and exchange. However, it is also known for the excessive information redundancy in its representation. While various compression schemes have been proposed and some of them can support query processing over compressed files, it is usually inevitable to perform partial (or full) data decompression which is expensive and in some cases may dominate the query processing time. In this paper, we propose a new XML compression scheme based on the Sequitur compression algorithm. By organizing the compression result as a set of context free grammar rules, the scheme supports efficient processing of XPath queries without decompression. The experimental results show that this scheme achieves comparable compression ratio as gzip while its query processing time is among the best of existing algorithms. Copyright 2005 ACM.