# Data Generation for Testing Complex Queries

arXiv: https://arxiv.org/abs/2409.18821 · HTML: https://mlsanity.com/paper/2409.18821/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 27 Sep 2024.

## Built on by (most cited first, 2 of 2)

- [SpotIt: Evaluating Text-to-SQL Evaluation with Formal Verification](https://mlsanity.com/paper/2510.26840/) (2025)
- [High-Fidelity And Complex Test Data Generation For Google SQL Code Generation Services](https://mlsanity.com/paper/2504.17203/) (2025)

## Abstract (arXiv)

Generation of sample data for testing SQL queries has been an important task for many years, with applications such as testing of SQL queries used for data analytics and in application software, as well as student SQL queries. More recently, with the increasing use of text-to-SQL systems, test data is key for the validation of generated queries. Earlier work for test data generation handled basic single block SQL queries, as well as simple nested SQL queries, but could not handle more complex queries. In this paper, we present a novel data generation approach that is designed to handle complex queries, and show its effectiveness on queries for which the earlier XData approach is not as effective. We also show that it can outperform the state-of-the-art VeriEQL system in showing non-equivalence of queries.