volume_mute

Natural Language to SQL System

publish date: 2026/06/11 07:14:53.963753 UTC

volume_mute

You need to design an architecture for a system that accepts natural language commands from users and translates them into SQL database queries. The system must parse the natural language input, identify the entities and relationships, and generate a valid SQL statement. Which of the following architectural patterns is most appropriate as the primary pattern, and which components would it include?

Pipe and Filter (or language processing system architecture) - with pipeline stages: Lexical Analysis → Syntactic Analysis → Semantic Analysis → SQL Code Generation, sharing a Symbol Table and Syntax Tree via a central repository

Client-Server - with the user as the client and the database as the server; no parsing components needed

MVC - with the natural language input as the Model, SQL as the View, and the database as the Controller

Layered architecture - with natural language at the top layer and SQL at the bottom layer, with no transformation components

Correct Answer

Explanation

A natural language to SQL translator is a language processing system. The appropriate architecture is a Pipe and Filter pipeline: Lexical Analysis (tokenizes the natural language input) → Syntactic Analysis (parses sentence structure) → Semantic Analysis (identifies entities, attributes, and relationships in the database schema context) → SQL Code Generation (produces the SQL query). A shared Symbol Table and Syntax Tree act as a central repository connecting the pipeline stages. This is the standard language processing system architecture applied to a new domain.

Reference

Software Engineering, Ian Sommerville, 10th edition

Quizzes you can take where this question appears

Architecture Design: Exercises