Type Review Framework

When reviewing a schema, apply the same checklist to every column. This catches type bugs before they reach production. The cost of fixing a type in design is a one-line change. The cost of fixing a type in production is a migration, a backfill, and a data reconciliation. The Checklist Common Type Smells The Operational Cost of Getting It Wrong Type changes on large tables are expensive. Changing a column from INT to BIGINT on a 1-billion-row table in PostgreSQL requires a full table rewrite. In Snowflake, type changes are metadata-only for some types (VARCHAR widening) but require a rewrite for others (INT to BIGINT). Always check your platform's ALTER TABLE behavior before declaring a type in production.