RapidMatch

Abstract

A subgraph query searches for all embeddings in a data graph that are identical to a query graph. Two kinds of algorithms, either graph exploration based or join based, have been developed for processing subgraph queries. Due to algorithmic and implementational differences, join-based systems can handle query graphs of a few vertices efficiently whereas exploration-based approaches typically process up to several tens of vertices in the query graph. In this paper, we first compare these two kinds of methods and prove that the complexity of result enumeration in state-of-the-art exploration-based methods matches that of the worst-case optimal join. Furthermore, we propose RapidMatch, a holistic subgraph query processing framework integrating the two approaches. Specifically, RapidMatch not only runs relational operators such as selections and joins, but also utilizes graph structural information, as in graph exploration, for filtering and join plan generation. Consequently, it outperforms the state of the art in both approaches on a wide range of query workloads.

GuP: Fast Subgraph Matching by Guard-based Pruning

“…On vertex-labeled graphs, subgraph containment algorithms [4,6,13] take a set of data graphs and find ones with at least one embedding of a query graph, and subgraph matching algorithms find all embeddings in a single data graph. Approaches based on join operations [1,28,37] are mainly used for subgraph homomorphism-based subgraph matching, which allows duplicate assignments of query vertices to the same data vertex. In contrast, subgraph isomorphism-based subgraph matching, the focus of this paper, prohibits it.…”

“…Table 1 lists the symbols used in this paper. We focus on vertexlabeled simple undirected graphs, similarly to previous studies [3,8,14,15,20,[35][36][37][38]. Note that our method can easily adapt to other kinds of graphs, such as directed graphs and edge-labeled graphs.…”

“…Comparison with Failing Set-based Pruning. Failing set-based pruning proposed in DAF [14] is one of the backjumping [32] methods and is popular in the database community [20,35,37]. Although both DAF and GuP exploit nogoods for pruning, GuP is more effective for two reasons.…”

“…However, theoretically analyzing their contribution is difficult because of their sensitivity to input graphs. Thus, following previous studies [14,20,35,37], we experimentally evaluate it in Section 4.…”

“…of query vertex 𝑢 𝑖 to data vertex 𝑣 𝑗 , there exists full embedding 𝑀 = {(𝑢 0 , 𝑣 1 ), (𝑢 1 , 𝑣 4 ), (𝑢 2 , 𝑣 7 ), (𝑢 3 , 𝑣 10 ), (𝑢 4 , 𝑣 0 )}. Since subgraph matching is NP-hard [15] and computationally expensive for complex graphs, efficient methods have been studied for a long time [3,8,14,15,20,[35][36][37][38].…”