How can we evaluate how data systems handle user-defined functions?

#UDFBench introduces a new framework for benchmarking UDF performance across data engines, using data from the @OpenAIREGraph

It fills a critical gap by enabling comparison, bottleneck detection, and optimisation validation for procedural logic.

Awarded VLDB Best Experiment, Analysis & Benchmark Paper 2025.

🔗 https://www.openaire.eu/udfbench-a-new-benchmark-for-udf-queries

#VLDB2025

How can we measure how well data systems handle user-defined functions?

#UDFBench, offers new framework to test user-defined functions across data engines using data from the @OpenAIREGraph

UDFBench fills a key gap: traditional benchmarks seldom evaluate custom procedural logic, but UDFBench lets you compare performance, spot bottlenecks, and validate optimisations.

👇