Reasoning on Property Graphs with Graph Generating Dependencies

TL;DR

This paper introduces Graph Generating Dependencies (GGDs) for expressing complex constraints on property graphs, along with algorithms for reasoning tasks and practical validation of data inconsistencies.

Contribution

It presents new algorithms for satisfiability, implication, and validation of GGDs, and demonstrates their effectiveness in detecting data inconsistencies.

Findings

Validation of GGDs can identify data inconsistencies.

Algorithms for reasoning about GGDs are computationally feasible.

GGDs are applicable to real-world graph data management.

Abstract

Graph Generating Dependencies (GGDs) informally express constraints between two (possibly different) graph patterns which enforce relationships on both graph's data (via property value constraints) and its structure (via topological constraints). Graph Generating Dependencies (GGDs) can express tuple- and equality-generating dependencies on property graphs, both of which find broad application in graph data management. In this paper, we discuss the reasoning behind GGDs. We propose algorithms to solve the satisfiability, implication, and validation problems for GGDs and analyze their complexity. To demonstrate the practical use of GGDs, we propose an algorithm which finds inconsistencies in data through validation of GGDs. Our experiments show that even though the validation of GGDs has high computational complexity, GGDs can be used to find data inconsistencies in a feasible execution time on both synthetic and real-world data.