-
Notifications
You must be signed in to change notification settings - Fork 36
Description
Summary
When using Run File as Workflow, the Databricks VS Code extension generates (or checks for) a workflow wrapper file.
However, when synced/uploaded, Databricks Workspace automatically converts the file into a notebook, stripping its file extension (.py, .ipynb, etc.).
The extension later queries the original path including the extension, which no longer exists in the workspace.
This results in:
RESOURCE_DOES_NOT_EXIST
This makes one‑click “Run File as Workflow” fail consistently.
How to Recognize
Open:
View → Output → Databricks Logs
You will see a call like:
workspace/get-status 404 RESOURCE_DOES_NOT_EXIST
The missing path will look like:
.../myfile-workflow-wrapper.py
or
.../myfile-workflow-wrapper.ipynb
…but in the Databricks Workspace UI, the file has been imported as a notebook, so its path has no extension.
Expected Behavior
The extension should detect that Databricks strips extensions from imported notebook files.
It should look up the actual canonical notebook path, not the file-based path with an extension.
Actual Behavior
Extension generates/checks wrapper file with full filename (incl. extension).
Databricks imports it as a notebook and drops the extension.
Extension queries the old path.
Workspace returns RESOURCE_DOES_NOT_EXIST.
“Run File as Workflow” fails.
Steps to Reproduce
Open any .py or .ipynb file in VS Code connected to Databricks.
Click Run File as Workflow.
Open Output → Databricks Logs.
Observe the 404 on workspace/get-status for the wrapper file path.