Dbt Deploy

IMPORTANT: profiles.yml MUST be placed inside the dbt project directory (alongside dbt_project.yml), NOT in ~/.dbt/. The snow dbt deploy command bundles profiles.yml from the project directory into the deployed project.

Workflow

Step 1: Validate Project

Goal: Ensure project is ready for deployment

Actions:

1. Check dbt_project.yml exists
2. Check profiles.yml exists and has no env_var() or password fields
3. Check models/ directory has at least one .sql file

If validation fails due to env_var() or password in profiles.yml / project files: This is a migration case. Load migrate/SKILL.md and run the migration workflow first, then return here to deploy the migrated project.

Step 2: Create Target Schema

Goal: Ensure target schema exists

CREATE SCHEMA IF NOT EXISTS <database>.<schema>;

Step 3: Check for External Access Requirements

Goal: Determine if the project needs external network access

When is external access needed? If the project needs to reach external hosts at runtime (e.g., to resolve packages, call APIs, etc.), it needs an External Access Integration (EAI) attached at deploy time.

Actions:

1. Determine whether the project requires external network access
2. If yes, find an available EAI:

   SHOW EXTERNAL ACCESS INTEGRATIONS;
   

3. Pick the integration that grants access to the required hosts

Step 4: Deploy Project

Goal: Upload project to Snowflake

Command (works for new projects AND updates):

snow dbt deploy <project_name> \
  --source <path_to_project> \
  --database <database> \
  --schema <schema> \
  --external-access-integration <integration_name>  # if project needs external network access

Parameters:

Example - Deploy without external packages:

snow dbt deploy MY_PROJECT --source /path/to/project --database DB --schema SCHEMA

Example - Deploy with external access:

snow dbt deploy MY_PROJECT --source /path/to/project --database DB --schema SCHEMA \
  --external-access-integration MY_EAI

Example - Update (creates VERSION$2, VERSION$3, etc.):

# Same command! Just point to updated source
snow dbt deploy MY_PROJECT --source /path/to/updated_project --database DB --schema SCHEMA \
  --external-access-integration MY_EAI

Step 5: Verify Deployment

Goal: Confirm project was deployed with correct version

snow dbt list --in schema <schema> --database <database>

--database defaults to the connection's database if omitted. --in schema defaults to all schemas if omitted.

Check versions:

SHOW VERSIONS IN DBT PROJECT <database>.<schema>.<project_name>;

Stopping Points

• ⚠️ Step 1: If profiles.yml has invalid fields

Output

Deployed dbt project in Snowflake, ready for execution.

Next Steps

After deployment, load execute/SKILL.md to run models.

Important: If you fixed an incremental model (changed is_incremental() logic, unique key, or strategy), you MUST execute with --full-refresh to rebuild the table from scratch. A normal run only processes new rows and won't fix data built by the old broken logic.