dbt¶

Horizon Catalog supports both dbt deployment types. Select your deployment type to get started:

Supported deployments¶

• dbt Cloud
• dbt Core (open source)

How the connector works¶

The dbt connector enriches the data sources already in Horizon Catalog with metadata from your dbt project: model-to-model lineage for any connected warehouse, and model and column descriptions for non-Snowflake warehouses such as BigQuery or Amazon Redshift. For Snowflake, Horizon Catalog uses the native object descriptions instead. The connector doesn’t ingest data; it layers dbt documentation and lineage on top of an existing warehouse data source.

Horizon Catalog reads the build artifacts that dbt produces rather than connecting to dbt itself. In the connector wizard, you select your dbt type (dbt Cloud or dbt Core) up front. Select the setup guide for your deployment to see the prerequisites and connection-form fields specific to it.

Requirements¶

To connect dbt to Horizon Catalog, you need the following:

• A warehouse data source that is already connected in Horizon Catalog. The dbt connector attaches to it so that dbt models map to the corresponding warehouse tables and views.
• A built dbt project. Horizon Catalog never needs access to dbt directly, but it does need the build artifacts that dbt produces (for dbt Core) or access to a dbt Cloud job that produces them (for dbt Cloud).

Choosing between dbt Cloud and dbt Core¶

Limitations¶

Each dbt connection is associated with a single dbt project (one dbt Cloud job, or one set of dbt Core artifacts). Adding dependent dbt projects to the same connection so Horizon Catalog can stitch cross-project lineage isn’t available yet. Connect each project as its own connection.