Combining Tree Structure Indexes with Structural Indexes in Query Evaluation on XML Data

Abstract

There are a variety of structural indexes which have been proposed to speed up path expression queries over XML data. They usually work by partitioning nodes in the data graph into equivalence classes and storing equivalence classes as index nodes. The size of a structural index is never larger than the size of the data graph. In the literature it is not always mentioned that the basic structure of XML document is tree-structure. In prior work [1], we introduce and describe a new improved approach for query evaluation on XML data. We consider the data graph of an XML data as the union of the basic tree and the link graph. The basic tree is indexed, that improves the query evaluation more efficiently. In this paper, we introduce and describe a new approach combining two technics: structural- and tree structure indexes. The data graph is simulated by a strong 1-index, in which the basic tree structure remains. Moreover, tree structure index can be built on the new structural index in linear complexity with efficient algorithms. Our experiments show that the new combinational approach is more efficient than we just apply tree structure or structural indexes separately.