Triple Code: An Efficient Labeling Scheme for Query Answering in XML Data

Abstract

Labeling schemes play crucial roles in the processing of XML queries. Designing labeling scheme that can support various common structural relationship queries (such as PC, AD, SR, DO, and LCA) among nodes, has received much research attention. Although there have existed a variety of labeling schemes such as prefix-based labeling, interval-based labeling and integer-based labeling and as well as their variants, these labeling schemes show weakness in one aspect or another. Interval-based labeling schemes can answer AD, PC and DO queries efficiently and can be created fast, but they can't process SR and LCA. Prefix-based labeling and integer-based labeling can process all of common structural relationship queries, but they are far less efficient than interval-based labeling for AD and DO queries. In this paper, we propose a new interval-based labeling scheme, called Triple code, which has a good performance on construction time, storage cost and query cost. In addition, to optimize the performance of LCA computation in XML data with big depth, we propose iTriple code (Triple code with information of ancestor). Experiment results demonstrate the benefits of our proposed labeling schemes over the previous approaches.