Automated Database Testing via LLM-Synthesized SQL Features

Suyang Zhong, Manuel Rigger · 2025

Read Automated Database Testing via LLM-Synthesized SQL Features by Suyang Zhong, Manuel Rigger (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-02012.txt.

Description

Suyang Zhong and Manuel Rigger use large language models to find bugs in database systems. Such systems are tested by tools that write pairs of queries that should give the same answer. When the answers differ, a bug is found. Older tools rely on query writers coded by hand, so they cover only a small set of query features. Their 2025 tool is called ShQveL.

ShQveL asks a language model to fill gaps in query sketches, then folds the new pieces into the test tool. That widens the range of features the tool can try. The team ran it on five database systems and found fifty-five new bugs. Fifty of those bugs were fixed soon after the team reported them.

Catalog record

CREATORSuyang Zhong, Manuel Rigger
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE79 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-02012

Browse the full library · About the Tower of Babel