Skip to content

Adapting the kit

  1. Replace raw_customers.csv and raw_orders.csv with synthetic fixtures that reflect your domain, or replace the seed references with declared warehouse sources.
  2. Keep source cleanup and type casting in models/staging/; keep business logic and aggregates in models/marts/.
  3. Add tests for key constraints and business rules in models/schema.yml. Add relationships tests when you have a defined orphan-record policy.
  4. Set model owners, freshness expectations, and alert routing to match your team's operational needs.
  5. Review the generated SQL and warehouse cost before deploying. Set sensible warehouse sizing and job schedules in your orchestrator.

Troubleshooting

  • Profile not found: ensure the file is at ~/.dbt/profiles.yml and its profile key is reliable_pipeline.
  • Missing environment variable: export the required SNOWFLAKE_* value in the same shell that runs dbt.
  • Permission denied: confirm the role can use the warehouse and create objects in the selected database and schema.
  • Seed parse/type issues: verify CSV headers match the model column names and that dates and amounts use valid formats.