dbt-commands

Use dbt's model selection syntax to target specific models by name, tag, folder, or pattern. For example, `dbt run -s dim_customers` runs a single model, while `dbt run -s tag:daily` runs all models tagged 'daily'. This lets you execute only the models you need without running the full build.

Use the `+` selector to include upstream dependencies: `dbt run -s +dim_customers` runs the model plus everything it depends on. The `dbt build` command also respects dependencies and executes in the correct order, handling both models and tests.

Run `dbt compile` to generate compiled SQL and identify Jinja macro errors without executing models. Check the `target/compiled` folder for the rendered SQL, review the error message for template syntax issues, and use `dbt debug` to verify project configuration and database connectivity.

Yes, use model and tag selection with `dbt test`: `dbt test -s dim_customers` tests only that model, and `dbt test -s tag:critical` runs tests for all models with the critical tag. This narrows test execution to relevant subsets of your data pipeline.

Use `--profiles-dir` to specify credential location, `--target` to select environment, `--threads` to control parallelism, and `--fail-fast` to stop on first error. Combine with selection syntax like `dbt run -s tag:prod_only` to control exactly which models execute in production.

Run `dbt docs generate` to create documentation artifacts, then `dbt docs serve` to view the generated docs locally. This exposes your model lineage, column descriptions, and test coverage—useful for understanding data pipeline dependencies and documentation.