The databricks bundle bind command does not support Unity Catalog resource types such as catalogs and external_locations. This makes it impossible to bind pre-existing UC assets to bundle resource definitions, so DAB always attempts to create new resources rather than manage
the existing ones.
Configuration
Steps to reproduce the behavior
- Define an external_location or Unity Catalog resource in a bundle YAML file.
- Deploy the bundle: databricks bundle deploy --target dev
- Attempt to bind the resource to an existing asset: databricks bundle bind external_location
- Observe that the bind subcommand does not accept external_location or catalog as a valid resource type.
Expected Behavior
databricks bundle bind external_location should work analogously to binding jobs or pipelines, allowing the bundle to manage a pre-existing UC external location instead of failing.
The same should apply to catalog resources.
Actual Behavior
The bind subcommand does not recognise external_location (or catalog) as a supported resource type. The CLI either returns an error.
This causes conflicts when the external location or catalog already exists in the workspace.
OS and CLI version
- OS: Windows 11 Pro
- Databricks CLI version: Databricks CLI v0.295.0
The databricks bundle bind command does not support Unity Catalog resource types such as catalogs and external_locations. This makes it impossible to bind pre-existing UC assets to bundle resource definitions, so DAB always attempts to create new resources rather than manage
the existing ones.
Configuration
Steps to reproduce the behavior
Expected Behavior
databricks bundle bind external_location should work analogously to binding jobs or pipelines, allowing the bundle to manage a pre-existing UC external location instead of failing.
The same should apply to catalog resources.
Actual Behavior
The bind subcommand does not recognise external_location (or catalog) as a supported resource type. The CLI either returns an error.
This causes conflicts when the external location or catalog already exists in the workspace.
OS and CLI version