# A Survey on Text-to-SQL Parsing: Concepts, Methods, and Future Directions

**Authors:** Bowen Qin, Binyuan Hui, Lihan Wang, Min Yang, Jinyang Li, Binhua Li, Ruiying Geng, Rongyu Cao, Jian Sun, Luo Si, Fei Huang, Yongbin Li  
**arXiv:** [2208.13629](https://arxiv.org/abs/2208.13629)  
**Date:** 2022-08-29  
**Summary page:** https://www.summarizepaper.com/en/paper/2208.13629v1/

## Key points

- Text-to-SQL parsing converts natural language questions into structured query language based on evidence from relational databases
- Early systems required significant human engineering and user interactions, but deep neural networks have revolutionized the field
- Deep neural networks introduce neural generation models that automatically learn the mapping function from NL questions to SQL queries
- Large pre-trained language models have elevated the state-of-the-art in text-to-SQL parsing
- The survey categorizes text-to-SQL parsing corpora as single-turn and multi-turn datasets, providing an overview of pre-trained language models and existing methods
- Challenges in the field are highlighted, along with potential future directions for advancing text-to-SQL parsing technology
- The paper serves as a valuable resource for researchers and practitioners interested in understanding text-to-SQL parsing concepts, methods, and future prospects within deep learning approaches

## Summary

In their paper titled "A Survey on Text-to-SQL Parsing: Concepts, Methods, and Future Directions," authors Bowen Qin, Binyuan Hui, Lihan Wang, Min Yang, Jinyang Li, Binhua Li, Ruiying Geng, Rongyu Cao, Jian Sun, Luo Si, Fei Huang and Yongbin Li delve into the essential and challenging task of text-to-SQL parsing. The primary objective of text-to-SQL parsing is to convert natural language (NL) questions into structured query language (SQL) based on evidence provided by relational databases. Early text-to-SQL parsing systems made significant progress within the database community but required substantial human engineering and user interactions. However, in recent years deep neural networks have revolutionized this field by introducing neural generation models that automatically learn a mapping function from NL questions to SQL queries. Additionally,<nl>large pre-trained language models have elevated the state-of-the-art in this field.</nl> This survey offers a comprehensive review of deep learning approaches for text-to-SQL parsing. The authors categorize text-to-SQL parsing corpora as single-turn and multi-turn datasets and provide a systematic overview of pre-trained language models and existing methods for text-to-SQL parsing. Furthermore,<nl>the paper highlights the challenges faced in this domain</nl> and explores potential future directions for advancing text-to-SQL parsing technology. Overall,<nl>this survey serves as a valuable resource for researchers and practitioners interested in understanding the concepts,</nl> methods,and future prospects of text-to-SQL parsing within the context of deep learning approaches.