Optimizing SPARQL queries in a distributed graph database

A non-transitory computer readable storage medium having recorded thereon a computer program that when executed by a computer causes the computer to implement a method for reducing an overhead of calls sent by a query engine to a storage engine, the query engine generating at least one graph of operators corresponding to a query execution plan of a SPARQL (SPARQL Protocol and RDF Query Language) query over a Resource Description Framework (RDF) graph, the RDF graph being distributed over different physical locations and having a partitioning into two or more subgraphs, the partitioning being not available to the method, the query execution plan being a sequence of steps to access data in a SQL (Structured Query Language) relational database management system, the method comprising:

• obtaining, by the query engine, a graph of operators executable by the query engine, each executable operator being in correspondence with at least one respective call for querying the storage engine, the obtained graph of operators including a plurality of basic operators, at least two of the basic operators of the obtained graph being of a first type each configured to find RDF triples of the RDF graph that match a respective basic graph pattern without knowing or imposing the partitioning of the RDF graph;
• and
• reducing data transmitted between the query engine and the storage engine by:
• identifying, by the query engine, a group of operators among the at least two basic operators of the obtained graph of operators which are of the first type such that respective basic graph patterns of the group of operators have same subject and/or predicate and/or object,
• replacing the identified group of operators in the obtained graph of operators by an equivalent operator,
• querying the storage engine with a single call corresponding to the equivalent operator,
• finding, based on the querying of the storage engine, RDF triples of the RDF graph that match the respective basic graph patterns of the group of operators, and
• accessing, using the RDF triples, data in the SQL relational database management system.

A system comprising:

• a processor coupled to a memory, the memory having recorded thereon a computer program for reducing an overhead of calls sent by a query engine to a storage engine, the query engine generating at least one graph of operators corresponding to a query execution plan of a SPARQL (SPARQL Protocol and RDF Query Language) query over a Resource Description Framework (RDF) graph, the RDF graph being distributed over different physical locations and having a partitioning into two or more subgraphs, the partitioning being not available to the method, the query execution plan being a sequence of steps to access data in a SQL (Structured Query Language) relational database management system, that when executed by the processor causes the processor to be configured to:
• obtain, by the query engine, a graph of operators executable by the query engine, each executable operator being in correspondence with at least one respective call for querying the storage engine, the obtained graph of operators including a plurality of basic operators, at least two of the basic operators of the obtained graph being of a first type each configured to find RDF triples of the RDF graph that match a respective basic graph pattern without knowing or imposing the partitioning of the RDF graph, and
• reduce data transmitted between the query engine and the storage engine by the processor being configured to:
• identify, by the query engine, a group of operators among the at least two basic operators of the obtained graph of operators which are of the first type such that respective basic graph patterns of the group of operators have same subject and/or predicate and/or object,
• replace the identified group of operators in the obtained graph of operators by an equivalent operator,
• query the storage engine with a single call corresponding to the equivalent operator,
• find, based on the querying of the storage engine, RDF triples of the RDF graph that match the respective basic graph patterns of the group of operators, and
• access, using the RDF triples, data in the SQL relational database management system.