UDFB ench : A Tool for Benchmarking UDF Queries on SQL Engines
Yannis Foufoulas, Theoni Palaiologou, Alkis Simitsis · 2025
User-defined functions (UDFs) extend SQL with functional capabilities. However, integrating UDFs efficiently into a data engine is a challenge, mainly due to the mismatch of the UDF execution and SQL processing environments. Several techniques have been proposed to deal with this challenge spanning a variety of logical and physical optimizations, and architecture designs. These however are not evident to the casual user, who would benefit from knowing the pros and cons of the available solutions. Towards this direction, we present UDFBench, a benchmarking toolkit for running seamlessly UDF queries (queries with UDFs) on a variety of SQL engines and presenting to the user execution results accompanied with highlights of the main reasons driving the query performance. UDFBench comprises a set of queries and UDFs of varying complexity based on a real-world application and data. Our current implementation has been deployed on four popular database engines with different characteristics. Our pluggable design enables extending UDFBench to additional systems and workloads.