System Guides
Databases & DWH
Pragmatic integration for SQL databases and data warehouses
Typical systems
- PostgreSQL, MySQL, MSSQL
- data warehouses (e.g. Snowflake, BigQuery, Redshift)
- reporting-oriented SQL backends
Pragmatic first
If direct access is possible
- start with the custom database approach
- validate first read-only use cases
- deliver fast answers on structured data questions
If tighter control is needed
- use Dataset Manager (SQL)
- leverage automatic schema introspection
- explicitly allow/limit fields
- define stable SQL snippets as functions
If direct access is not possible
- provide scheduled exports (CSV/SQL dumps)
- start with Dataset Manager or data-pool-adjacent workflows
Which path when?
- operational live queries -> direct DB access
- repeatable analytics -> Dataset Manager
- constrained infrastructure -> export-first approach
Practical 3-step start
- pick 1-2 tables with clear business questions
- apply read-only setup and query limits
- validate responses with business owners before scaling