On Defining SPARQL with Boolean Tensor Algebra

Abstract: The Resource Description Framework (RDF) represents information as subject-predicate-object triples. These triples are commonly interpreted as a directed labelled graph. We propose an alternative approach, interpreting the data as a 3-way Boolean tensor. We show how SPARQL queries - the standard queries for RDF - can be expressed as elementary operations in Boolean algebra, giving us a complete re-interpretation of RDF and SPARQL. We show how the Boolean tensor interpretation allows for new optimizations and analyses of the complexity of SPARQL queries. For example, estimating the size of the results for different join queries becomes much simpler.