Graph database systems (GDBs) allow efficiently storing and retrieving graph data, and have become the critical component in many applications, e.g., knowledge graphs, social networks, and fraud detection. It is important to ensure that GDBs operate correctly. Logic bugs can occur and make GDBs return an incorrect result for a given query. These bugs are critical and can easily go unnoticed by developers when the graph and queries become complicated. Despite the importance of GDBs, logic bugs in GDBs have received less attention than those in relational database systems.

In this paper, we present Grand, an approach for automatically finding logic bugs in GDBs that adopt Gremlin as their query language. The core idea of Grand is to construct semantically equivalent databases for multiple GDBs, and then compare the results of a Gremlin query on these databases. If the return results of a query on multiple GDBs are different, the likely cause is a logic bug in these GDBs. To effectively test GDBs, we propose a model-based query generation approach to generate valid Gremlin queries that can potentially return non-empty results, and a data mapping approach to unify the format of query results for different GDBs. We evaluate Grand on six widely-used GDBs, e.g., Neo4j and HugeGraph. In total, we have found 21 previously-unknown logic bugs in these GDBs. Among them, developers have confirmed 18 bugs, and fixed 7 bugs.