Independence of Containing Patterns Property and Its Application in Tree Pattern Query Rewriting Using Views

• Authors: Junhu Wang;Jeffrey Xu Yu;Chengfei Liu
• Venue: World Wide Web
• Year: 2009

We show that several classes of tree patterns observe theindependence of containing patterns property, that is, ifa pattern is contained in the union of several patterns, then it iscontained in one of them. We apply this property to two relatedproblems on tree pattern rewriting using views. First, given viewV and query Q, is it possible for Q tohave an equivalent rewriting using V which is the union oftwo or more tree patterns, but not an equivalent rewriting which isa single pattern? This problem is of both theoretical and practicalimportance because, if the answer is no, then, to find anequivalent rewriting of a tree pattern using a view, we should usemore efficient methods, such as the polynomial time algorithm of Xuand Özsoyoglu (2005), rather than try to find the union of allcontained rewritings (which takes exponential time in the worstcase) and test its equivalence to Q. Second, given a setS of views, we want to know under what conditions a subsetS' of S is redundant in the sense that forany queryQ, the contained rewritings ofQ using the views in S' are contained in thoseusing the views in S--S'. Solving this problemcan help us to, for example, choose the minimum number of views tobe cached, or better design the virtual schema in a mediated dataintegration system, or avoid repeated calculation in queryoptimization. For the first problem, we identify several classes oftree patterns for which the equivalent rewriting can be expressedas a single tree pattern. For the second problem, we presentnecessary and sufficient conditions for S to be redundantwith respect to some classes of tree patterns. For both problems weconsider extension to cases where there are rewritings using theintersection of multiple views and/or where a schema graph ispresent.