Representing Paths in Graph Database Pattern Matching

TL;DR

This paper introduces path multiset representations (PMRs) for graph databases, enabling efficient storage and processing of path queries, significantly improving time and space performance for complex path-based queries.

Contribution

The paper presents PMRs as a novel, succinct way to represent path multisets, along with analysis of their minimization, equivalence, and application benefits in query execution.

Findings

PMRs enable exponential compression of path multisets.

PMRs facilitate faster query plan execution with reduced space.

PMRs are well-suited for complex regular path query extensions.

Abstract

Modern graph database query languages such as GQL, SQL/PGQ, and their academic predecessor G-Core promote paths to first-class citizens in the sense that paths that match regular path queries can be returned to the user. This brings a number of challenges in terms of efficiency, caused by the fact that graphs can have a huge amount of paths between a given node pair. We introduce the concept of path multiset representations (PMRs), which can represent multisets of paths in an exponentially succinct manner. After exploring fundamental problems such as minimization and equivalence testing of PMRs, we explore how their use can lead to significant time and space savings when executing query plans. We show that, from a computational complexity point of view, PMRs seem especially well-suited for representing results of regular path queries and extensions thereof involving counting, random sampling, unions, and joins.