We tackle the containment problem for conjunctive queries with negation, which takes two queries q 1 and q 2 as input and asks if q 1 is contained in q 2. A general approach for solving this problem consists of considering all completions of q 1 (intuitively these completions represent all canonical databases that satisfy q 1) and checking if each completion yields the same answer on q 2. Since the total number of completions of q 1 is exponential in the size of q 1, several proposals have aimed at reducing the number (and the size) of the completions checked. In this paper, we propose a unifying framework for comparing algorithms following this general approach and define two kinds of heuristics for exploring the space of completions. Combining these heuristics with both classical kinds of traversals, i.e., depth-first and breadth-first, we obtain four algorithms that we compare experimentally with respect to running time on difficult instances of the containment problem. © 2011 Springer-Verlag Berlin Heidelberg.
CITATION STYLE
Mohamed, K. B., Leclère, M., & Mugnier, M. L. (2011). A theoretical and experimental comparison of algorithms for the containment of conjunctive queries with negation. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 6860 LNCS, pp. 466–480). https://doi.org/10.1007/978-3-642-23088-2_35
Mendeley helps you to discover research relevant for your work.