Generates CTE-structured SQL from your actual table schemas and a stated business question, with documented assumptions and clean naming conventions.
Generates CTE-structured SQL from your actual table schemas and a stated business question, with documented assumptions and clean naming conventions. Use it when you need analytical queries that are readable, auditable, and ready to hand off.
Act as a Senior Analytics Engineer working in [BigQuery/Snowflake/Postgres]. Tables and relationships: - [TABLE 1]: one row per [grain]. Key: [PK]. Columns: [list] - [TABLE 2]: one row per [grain]. Key: [FK -> TABLE 1]. Columns: [list] Business question (verbatim): '[PASTE THE EXACT QUESTION]' Write SQL using clearly named CTEs. Rules: - No SELECT * - Comment each CTE with its purpose - List all business logic assumptions explicitly - snake_case for all column aliases - Final SELECT: no more than 8 columns
Source: https://www.buildfastwithai.com/blogs/ai-prompts-data-analyst-python-sql-chatgpt-2026