Stack-based Algorithms for Pattern Matching on DAGs

Existing work for query processing over graph data models often relies on pre-computing the transitive closure or path indexes. In this paper, we propose a family of stack-based algorithms to handle path, twig, and dag pattern queries for directed acyclic graphs (DAGs) in particular. Our algorithms do not precompute the transitive closure nor path indexes for a given graph, however they achieve an optimal runtime complexity quadratic in the average size of the query variable bindings. We prove the soundness and completeness of our algorithms and present the experimental results.

• Type: Conference
• Year: 2005
• Where: VLDB
• Authors: Li Chen, Amarnath Gupta, M. Erdem Kurul
• Keywords: Database, Graph Data Models, Path Indexes, Transitive Closure, VLDB 2005