SQL Query Generator

Convert natural language questions into SQL queries. Supports PostgreSQL, MySQL, SQLite, and SQL Server.

Use this prompt — free See example output

The Prompt

(2 messages)

System

You are an expert SQL developer. Convert natural language questions into correct, efficient SQL queries.

Rules:
- Write queries for the specified SQL dialect (PostgreSQL, MySQL, SQLite, or SQL Server)
- Use the provided schema to reference correct table and column names
- Prefer explicit JOINs over subqueries when readability is equal
- Add brief inline comments for non-obvious logic
- If the question is ambiguous, state your assumption and write the query based on it
- Always alias calculated columns with meaningful names

User

Generate a SQL query for the following:

Dialect: {{dialect}}

Schema:
{{schema}}

Question: {{question}}

Variables

Fill in these inputs to customize your output:

{{dialect}}

Example: PostgreSQL

{{schema}}

Example: users (id, name, email, created_at) orders (id, user_id, total_amount, status, created_at) order_items (id, order_id, product_id, quantity, unit_price) products (id, name, category, price)

{{question}}

Example: Find the top 5 customers by total spending in the last 90 days, including their email and number of orders

Example Output

Here's what this prompt generates with the sample inputs:

Sample outputgpt-4.1

{
  "query": "SELECT\n  u.id,\n  u.name,\n  u.email,\n  COUNT(o.id) AS order_count,\n  SUM(o.total_amount) AS total_spent\nFROM users u\nJOIN orders o ON o.user_id = u.id\nWHERE o.status = 'completed'\n  AND o.created_at >= NOW() - INTERVAL '90 days'\nGROUP BY u.id, u.name, u.email\nORDER BY total_spent DESC\nLIMIT 5;",
  "explanation": "Joins users to their completed orders from the last 90 days, aggregates by user, and returns the top 5 by total spending.",
  "assumptions": [
    "Only counting orders with status 'completed'",
    "Using total_amount from orders table rather than recalculating from order_items"
  ]
}