Multi Reachability Query Processing

There is a need to efficiently navigate on a large graph to find a certain piece of information by some structure constraints. Reachability query could be the main means for such a purpose and existing approaches studied an efficient primitive called reachability join. In this paper, we focus on processing the query containing multi reachability joins (R-joins). We studied the processing ability of the up-to-date alternative that extended the wellknown tree-specific method, namely, twig-stack join algorithm, to be applicable on DAGs. Due to the rapid increasing of the transitive closure size for graphs in stead of trees, however, we found the proposed dynamic programming approach that combines R-join primitives into an optimized plan based on cost tends to be superior over the alternative and more stable in terms of the density of the underneath DAGs than the alternative. According to the extensive experiments we’ve done, our method is faster than the existing work by at least two orders of magnitudes.