SgIndex: An Index Structure Supporting Multiple Graph Queries

Abstract

With the rise of social networks, traffic navigation and other fields, graph applications become increasing extensive. To improve query efficiency, indexes are built to manage large-scale graph data. However, these indexes cost large memory space and can only support one single graph operation. We propose a two-layered index structure SgIndex, in which the first layer stores subgraph information, and the second layer stores adjacency information to support multiple path operations and subgraph matching queries. We propose a subgraph matching algorithm based on path join, which completes subgraph matching by searching SgIndex twice. The experimental results show that SgIndex achieves better performance on path queries and subgraph matching than existing index structures, and reduces memory overhead.