Path Query Processing on Very Large RDF Graphs
Andrey Gubichev, Thomas Neumann · 2011
Finding the shortest path between two nodes in an RDF graph is a fundamental operation that allows to discover complex relationships between entities. In this paper we consider the path queries over graphs from a database perspective. We provide the full-fledge database solution to execute path queries over very large RDF graphs. We present low-level techniques to speed-up shortest paths algorithms, and a robust method to estimate selectivities of path queries. We perform extended experiments on several large RDF collections, including the UniProt collection, demonstrating that our approach outperforms the path query capabilities of modern systems by a large margin. 1.