No workspace file tree shown, cannot import local modules using databricks kernels

[DrMaphuse]

I can't seem to get proper access to my databricks workspace without running into a Cannot read properties of undefined (reading 'path')-Error.

My setup so far:

• I am using databricks-connect with the official db extension
• I can click on the orange Add databricks workspace to VSCode workspace-button and it seems to load my databricks workspace and somehow also connects it to a local clone of the same repo.
• I can see clusters, DBFS, secrets etc.,
• I can use my shared cluster as a python kernel (dbr 16.4, using databricks-connect 16.4)
• I CANNOT see the file tree of my workspace in the extension as shown in the README
• I CANNOT access the files and modules in my repo from within the kernel (imports do not work, os.listdir() shows only some hidden directories

[gbrueckl]

hi @DrMaphuse and sorry for the late response

may I ask which kernel you are using - the one from the Databricks Connection or the one from Databricks Power Tools?

by clicking Add databricks workspace to VSCode workspace we basically create an in-memory representation of your Databricks workspace within VSCode. You can work with it the very same way as with a local file system. If the folder does not show up in VSCode Explorer, please check your workspace settings - you should see something like this - a wsfs:/ folder:

you might also gather a lot more information by checking the VSCode output channel for paiqo.databricks-vscode