(A) queriable XML compression technique using a type inference engine

Abstract

Like HTML, many XML documents are resident on native file systems. Since XML data is irregular and verbose, the disk space and the network bandwidth are wasted compared to those of regularly structured data. To overcome the verbosity problem of XML data, the research on compressors for XML data has been conducted. However, some XML compressors do not support querying compressed data, while other XML compressors which support querying compressed data make the assumption for the total number of distinct tags in XML data in order to compress them and blindly encode data values using predefined encoding methods without considering the types of data values. Thus, the query performance on compressed XML data is degraded. This research proposes an XML compression technique which supports direct and efficient evaluations of queries on compressed XML data. This XML compression technique basically adopts an encoding method, called dictionary encoding, in order to encode each tag of XML data into a unique integer value, and applies diverse encoding methods for encoding data values depending on the inferred types of data values. Those inferred types are acquired by the type inference engine which distinguishes the appropriate types of data values of each element and attribute. Furthermore, this XML compression technique does not make any assumptions for the total number of distinct tags in XML data. This XML compression technique is implemented as an XML compressor, named XTIE, to evaluate the performance of this XML compression technique. Experimental results with real-life data sets show that XTIE achieves significant improvements on query performance for compressed XML data and reasonable compression ratios compared to other existing XML compressors. On the average, the query performance of XTIE is about 1.4 times better than that of an existing XML compressor and the compression ratio of XTIE is about 80%.