Graph Processing Using Relational Operators

December 20, 2025 · View on GitHub

Worst-Case Optimal Join (WCOJ) Algorithms and Factorized Query Processing

GraphflowDB and Modern Query Processing Techniques for GDBMSs, 1-hour lecture by Semih Salihoglu at Pinterest Labs (2021) [WCOJ] [FACT]
From joins to aggregates and optimization problems (slides), lecture by Dan Olteanu (2018) [WCOJ] [FACT]
From joins to aggregates and optimization problems, lectures series by Dan Olteanu (2018) [WCOJ] [FACT]
Factorised databases, slides, short lecture by Dan Olteanu (2017) [FACT]
SIGMOD 2020 tutorial: Optimal Join Algorithms Meet Top-k [WCOJ]
LIquid: Soul of a New Graph Database by Scott Meyer, presented at the University of Waterloo's DSG Seminar Series
- YouTube
Relational databases can handle graphs too! Experiences with optimizing the Umbra RDBMS for LDBC SNB BI (slides) by Altan Birler (Technische Universität München), presented at the 15th LDBC TUC

⚡️ Graph Processing in RDBMSs by Kangfei Zhao, Jeffrey Xu Yu (IEEE Data Eng. Bull., 2017)
Semantic Data Modeling, Graph Query, and SQL, Together at Last? by Jeff Shute, Colin Zheng, Romit Kudtarkar (CIDR 2026)

Papers from the University of Waterloo:

Other papers:

Adopting Worst-Case Optimal Joins in Relational Database Systems Michael Freitag et al. (VLDB 2020): WCOJs in the context of the Umbra compiled columnar RDBMS
Towards Multi-way Join Aware Optimizer in SAP HANA by Sungheun Wi et al. (VLDB 2020)
A Worst-Case Optimal Join Algorithm for SPARQL by Aidan Hogan et al. (ISWC 2019)

The case against specialized graph analytics engines (CIDR 2015), a paper showing how implementing graph algorithms (SSSP, PageRank, connected components) in T-SQL (Microsoft SQL Server) can outperform the implementations of dedicated graph analytics engines
In-database connected component analysis (ICDE 2020), a connected components implementation in SQL