BBTC: A New Update-aware Coding Scheme for Efficient Structure Join

The identification of ancestor-descendant or parent-child relationship between elements of XML documents plays a crucial role in efficient XML query processing. One of the popular methods for performing this task is to code each node in the XML document by traversing its nodes. However, the main problems of existing approaches are that they either lack the ability to support XML document update or need huge storage space. This paper proposes a novel coding scheme called Blocked Binary-Tree Coding scheme (BBTC) by taking the issues of identification, easy update and low storage cost into account. BBTC identifies the ancestor-descendant relationship in constant time. For the update, only a few simple operations for the affected document elements are needed. More importantly, for BBTC, this paper proposes a structure join algorithm BDC based on Bucket Divide and Conquer. BDC not only accelerates structure join dramatically when the input element sets are neither sorted nor indexed, but also can be applied to other coding schemes. The extensive experiments show that both the coding scheme BBTC and BDC significantly outperform the existing studies.