Optimizing SPARQLeR using short circuit evaluation of filter clauses
Abstract
The discovery of semantic relationships between two entities is an area of interest and study in the Semantic Web. Regarding the data in an RDF store as a graph allows the utilization of path searches, but with large ontologies exhaustive searches are not practical. SPARQLeR, an extension of SPARQL, sought to utilize regular expressions in the search and construction of a path which helped avoid exhaustive searches by limiting the search space during evaluation. Still, due to the filter constraint evaluation occurring after a path was constructed, the actual search space of the beginning and ending resources was not being limited. To address this problem, an optimization is proposed that uses a form of short circuit evaluation of the filter constraints in a SPARQL query to limit the search space as early as possible to avoid costly path searches between two nodes.