Abstract
|
Distributed system consists of multiple databases which are interconnected via a communication network. While these systems have long been used by organizations, researchers have always attempted to solve the problem of making coordination between databases with different structures. This problem may be solved by XML language which can be easily converted to any format. With the increased volume of distributed documents, optimal query processing has become crucially important. XML queries consist of a series of elements which are interconnected under a tree structure. Therefore, finding the pattern between query and document is the central core of query processing. While many methods have been proposed for query processing, they all have the problem of processing the nodes which are not involved in the final answer. Consequently, these methods tend to waste time by processing useless nodes.
The present paper proposes a new method for query processing. This method processes the nodes which are definitely involved in final answer. In contrast to other methods, this method works with a lot of indexes and efficiently answers different kinds of query. We examined the efficiency of this method using famous databanks such as DBLP, TreeBank, XMark as well as balanced and unbalanced random databanks. We also tested simple query, single-branch query and multi-branch query with and without extraction point. The results indicated that the proposed method was more efficient than the existing popular methods in terms of the number of processed nodes, used memory and execution time
|