Research on SPARQL Semantic Query Technology Based on Knowledge Hybrid Storage

Yabiao Zhang, Bihui Yu, Jun Wang

Abstract: At present, knowledge graphs are usually expressed in RDF and given semantics. SPARQL is a semantic query language recommended by W3C. With the development of the Semantic Web technology, the current different storage modes for knowledge management mainly have the incompatibility of support for massive knowledge storage and semantic query standards. For example, Neo4j cannot use SPARQL for semantic query. Aiming at the low efficiency of technology fusion query conversion under the knowledge storage hybrid storage mode, this paper introduces the Apache TinkerPop graph calculation framework to build a bridge between RDF graphs and attribute graphs, and proposes a method for converting SPARQL queries to Gremlin attribute graph queries. Realize the semantic query ability under the massive knowledge storage, and make up for the shortcomings of different storage modes. In addition, TinkerPop must use a certain constant key value for binding query to expand to meet the completeness requirement of knowledge query. Finally, the experiment was verified by building a microservice API with Spring Boot.