Synthesis of SQL Queries from Narrations

Abstract

Structured Query Language (SQL) remains a standard language used in Relational Database Management Systems (RDBMSs), and has found applications in healthcare (patient registries), businesses (inventories, trend analysis), military, and education, etc. Although, SQL statements are English-like, the process of writing SQL queries is often problematic for non-technical end-users. To address this problem, a tool called Narrations-2-SQL is developed to allow an end-user to specify a query in natural language. Narrations-2-SQL is a desktop application that uses a Jumping Finite Automaton (JFA) - a type of Finite Machine for translating natural language descriptions into SQL queries, execute the queries, and provide a feedback to a user. An experimental evaluation was performed on 204 crowdsourced queries in natural language from the XNorthwind DB. Our results show an accuracy of 88%. To get the users' perceptions of this study, we carried out a survey on 167 end-users. Majority of the participants found Narrations-2-SQL to be very helpful, and agreed that it could be useful in industry. If implemented on a large scale, the tool may be helpful to many end-users in different domains.