The paper "gFOV: A Full-Stack SPARQL Query Optimizer & Plan Visualizer" by Yue Pang and others has been accepted by the CIKM 2023 Demo Track.

SPARQL is the standard query language for RDF data. A SPARQL query consists of basic graph patterns (BGPs), which are matched onto the data graph, and graph pattern operators (such as UNION and OPTIONAL) defined on top of it that specify how to merge the matched results. Despite the prevalence of graph pattern operators in real-world SPARQL workloads, research on optimizing SPARQL queries with graph pattern operators has been limited. Therefore, we propose gFOV, a full-stack SPARQL query optimizer targeting both basic graph patterns and graph pattern operators for joint optimization. We introduce a novel BGP-based evaluation tree (BE-tree) plan representation that integrates the physical plan for basic graph patterns, directly accessing the RDF store, and the logical plan for graph pattern operators operating on existent results in memory. On top of it, we design a full-stack cost-based optimization scheme, combining logical and physical plan optimization, which surpasses the current state-of-the-art. In the demonstration, we provide an interactive interface to explain our optimization scheme and showcase its efficiency by visualizing changes in the query plan, allowing the audience to inspect and execute alternative plans.