RelSeeker: Relationship-based Query Language in a Graph Database for Social Networks

Abstract

Thanks to the current massive increase in data related to social activity, such as social networks and real-time recommendations, large amounts of related data are generated every second. Traditional relational databases are insufficiently able to handle this situation. A graph database is suitable for representing relationships in social activity because social networks are easily represented by graphs. Graph databases allow users to store data as nodes and relationships so that the data may be queried as a graph. However, no mechanisms exist that can effectively carry out queries regarding relationships, such as queries about properties of a relationship or relationships of relationships. To solve these problems, we introduce the concept of relationship-based query and develop a graph database query language called RelSeeker, which is based on Datalog. In RelSeeker, a user can create query rules using the properties of relationships, nodes, and relationships of relationships. Moreover, Datalog is suitable for describing relationships of relationships in the graph database. These key features of RelSeker enable effective relationship-based query processing on a graph database. The proposed query language RelSeeker is an extension of Datalog, which can deal with data structures and manipulates a graph database through creating, getting, and updating operations.