What are workspace files?
A workspace file is any file in the Azure Databricks workspace that is not a Databricks notebook. Workspace files can be any file type. Common examples include:
.pyfiles used in custom modules..mdfiles, such asREADME.md..csvor other small data files..txtfiles..whllibraries.- Log files.
Workspace files include files formerly referred to as "Files in Repos."
Important
Workspace files are enabled everywhere by default in Databricks Runtime version 11.2, but can be disabled by admins using the REST API. For production workloads, use Databricks Runtime 11.3 LTS or above. Contact your workspace administrator if you cannot access this functionality.
What you can do with workspace files
Azure Databricks provides functionality similar to local development for many workspace file types, including a built-in file editor. Not all use cases for all file types are supported. For example, while you can include images in an imported directory or repository, you cannot embed images in notebooks.
You can create, edit, and manage access to workspace files using familiar patterns from notebook interactions. You can use relative paths for library imports from workspace files, similar to local development. For more details, see:
- Workspace files basic usage
- Programmatically interact with workspace files
- Work with Python and R modules
- Manage notebooks
- File ACLs
Init scripts stored in workspace files have special behavior. You can use workspace files to store and reference init scripts in any Databricks Runtime versions. See Store init scripts in workspace files.
Note
In Databricks Runtime 14.0 and above, the the default current working directory (CWD) for code executed locally is the directory containing the notebook or script being run. This is a change in behavior from Databricks Runtime 13.3 LTS and below. See What is the default current working directory?.
Limitations
A complete list of workspace files limitations is found in Workspace files limitations.
File size limit
Individual workspace files are limited to 500 MB.
Users can upload file sizes up to 500 MB from the UI. The maximum file size allowed when writing from a cluster is 256 MB.
Databricks Runtime versions for files in Repos with a cluster with Azure Databricks Container Services
On clusters running Databricks Runtime 11.3 LTS and above, the default settings allow you to use workspace files in Repos with Azure Databricks Container Services (DCS).
On clusters running Databricks Runtime versions 10.4 LTS and 9.1 LTS, you must configure the dockerfile to access workspace files in Repos on a cluster with DCS. Refer to the following dockerfiles for the desired Databricks Runtime version:
See Customize containers with Databricks Container Service
Enable workspace files
To enable support for non-notebook files in your Databricks workspace, call the /api/2.0/workspace-conf REST API from a notebook or other environment with access to your Databricks workspace. Workspace files are enabled by default.
To enable or re-enable support for non-notebook files in your Databricks workspace, call the /api/2.0/workspace-conf and get the value of the enableWorkspaceFileSystem key. If it is set to true, non-notebook files are already enabled for your workspace.
The following example demonstrates how you can call this API from a notebook to check if workspace files are disabled and if so, re-enable them. To disable workspace files, set enableWorkspaceFilesystem to false with the /api/2.0/workspace-conf API.