On the query evaluation in document DBs

In this paper, we study the query evaluation in document databases. First, we show that a query represented in an XML language can be generally considered as a labeled tree, and the evaluation of such a query is in fact a tree embedding problem. Then, we propose a strategy to solve this problem, based on dynamic programming. For the ordered tree embedding, the proposed algorithm needs only O(vertical bar T vertical bar(.)vertical bar P vertical bar) time and O(vertical bar T vertical bar(.)vertical bar P vertical bar) space, where vertical bar T vertical bar and vertical bar P vertical bar stands for the numbers of the nodes in the target tree T and the pattern tree P, respectively. This computational complexity is better than any existing method on this issue. In addition, how to adapt this method to the general tree embedding is also discussed.