Parallel Construction of Knowledge Graphs from Relational Databases

Knowledge graphs have recently seen a wide range of applications in various domains. In many such applications data stored in relational databases constitutes an important source for the construction of knowledge graphs. R2RML is a mapping language that can be used to specify mappings from relational to RDF data, and so it naturally suits the purpose of knowledge graph construction from relational data. In this paper, we present Fingr, a concurrent dictionary aided parallel R2RML engine that achieves fine-grained parallelization at the database tuple level. Our experiments show that our prototypical system parallelizes well, and it yields better performance than existing R2RML engines.