Text-to-SQL aims to automate the process of generating SQL queries on a database from natural language text. In this work, we propose "SQLPrompt", tailored to improve the few-shot prompting capabilities of Text-to-SQL for Large Language Models (LLMs). Our methods include innovative prompt design, execution-based consistency decoding strategy which selects the SQL with the most consistent execution outcome among other SQL proposals, and a method that aims to improve performance by diversifying the SQL proposals during consistency selection with different prompt designs ("MixPrompt") and foundation models ("MixLLMs"). We show that \emph{SQLPrompt} outperforms previous approaches for in-context learning with few labeled data by a large margin, closing the gap with finetuning state-of-the-art with thousands of labeled data.
@article{arxiv.2311.02883,
title = {SQLPrompt: In-Context Text-to-SQL with Minimal Labeled Data},
author = {Ruoxi Sun and Sercan Ö. Arik and Rajarishi Sinha and Hootan Nakhost and Hanjun Dai and Pengcheng Yin and Tomas Pfister},
journal= {arXiv preprint arXiv:2311.02883},
year = {2023}
}