Distinguish staging work from downstream business logic in a dbt transformation layer.
Good staging removes source friction without deciding the business answer. A dbt staging model should make raw data legible and stable. That usually means renaming fields, casting types, trimming malformed values, standardizing null handling, and preserving row grain. It should not silently become the company's policy engine. What belongs in staging Source-specific clean-up belongs here: customerID becomes customer_id, text booleans become real booleans, timezone-naive timestamps get normalized, and duplicate ingestion rows get handled according to a source-aware rule. What does not belong in staging If the line of SQL answers a business question that stakeholders might argue about, it probably…
Sign up free — one personalized lesson every day, matched to your role and goals.
Already have an account? Sign in