基于自然语言处理的SQL语句生成研究

Research on SQL Statement Generation Based on Natural Language Processing

大数据时代下,数据的价值也随之提升,数据库作为海量数据的存储和管理工具,熟练掌握其查询语言至关重要。为了提升非专业数据库用户使用数据库的简捷度以及提高其数据库操作的效率和准确性,本研究提出了一种基于ChatGLM预训练模型的自然语言转SQL查询语句的方法,即通过分词化、词嵌入、模型训练和模型测试等流程完成文本转SQL研究。在由1020条SQL查询语句构建的数据集上的实验表明,该方法的F1指数达到89.8%,能够满足普通用户的查询需求,促进了数据库的普及和应用。

In the era of Big data,the value of data will also increase.As a storage and management tool for massive data,it is crucial to master its Query language skillfully.In order to improve the simplicity of non professional database users using the database and improve the efficiency and accuracy of their database operations,this paper proposes a natural language to SQL query statement method based on the ChatGLM pre training model,that is,to complete the text to SQL research through processes such as lexicalization,Word embedding,model training and model testing.The experiment on a dataset constructed from 1020 SQL query statements shows that the F1 index of this method reaches 89.8%,which can meet the query needs of ordinary users and promote the popularization and application of databases.