Natural Language to SQL Queries: A Review

Abstract

The relational database is the way of maintaining, storing, and accessing structured data but in order to access the data in that database the queries need to be translated in the format of SQL queries. Using natural language rather than SQL has introduced the advancement of a new kind of handling strategy called Natural Language Interface to Database frameworks (NLIDB). NLIDB is a stage towards the turn of events of clever data set frameworks (IDBS) to upgrade the clients in performing adaptable questioning in data sets. A model that can deduce relational database queries from natural language. Advanced neural algorithms synthesize the end-to-end SQL to text relation which results in the accuracy of 80% on the publicly available datasets. In this paper, we reviewed the existing framework and compared them based on the aggregation classifier, select column pointer, and the clause pointer. Furthermore, we discussed the role of semantic parsing and neural algorithm’s contribution in predicting the aggregation, column pointer, and clause pointer. In particular, people with limited background knowledge are unable to access databases with ease. Using natural language interfaces for relational databases is the solution to make natural language to SQL queries. This paper presents a review of the existing framework to process natural language to SQL queries and we will also cover some of the speech to SQL model in discussion section, in order to understand their framework and to highlight the limitations in the existing models.