What you will be able to do
- Explain what a Databricks Git folder is and how it differs from a legacy Repo
- Configure Git credentials so that commits from Databricks are attributed to you
- Clone a remote repository into a Git folder from the workspace UI
- Open the Git dialog and name what each of its controls does
Key concept
Databricks Git folder — A workspace folder that is a clone of a remote Git repository. It tracks a branch of that repository, and you run Git operations on it (branch, commit, push, pull) from the workspace UI instead of from a local terminal.
1.From Repos to Git folders
The exam guide says 'Databricks Git Folders (formerly Databricks Repos)', and the documentation states it directly: 'Repos are now called Git folders.' A Git folder lets you develop notebooks and files in the workspace while following Git practice. You clone, push to, and pull from a remote repository. You create and manage branches, including merging, rebasing, and resolving conflicts. You also compare diffs visually when you commit.
The Git operations didn't change. The folder placement, the remote URL requirement and the supported asset types did. Repos could only sit at one fixed path. A Git folder can sit at any depth under /Workspace/Users/<user email>/, and you can have as many as you need. Older paths still resolve: /Repos and /Workspace/Repos point to the same folder, so existing %run references and job paths keep working. Databricks recommends creating new Git folders rather than Repos.
| Aspect | Legacy Repos | Git folders |
|---|---|---|
| How you create one | New > Repo | New > Git folder |
| Location | Fixed level: /Workspace/Repos/<user email>/<Repo name> | Any depth, e.g. /Workspace/Users/<user email>/level_1/level_2/level_3/<Git folder name> |
| Remote repository URL | Not required | Required |
| Asset types | Fewer supported | Can also hold assets such as Databricks SQL assets and MLflow experiments |
| Permissions | Separate Repos area | Same workspace folder permissions as other folders. Most Git operations require CAN MANAGE |
Checkpoint 1 of 5· Check yourself
A colleague wants to create a Git folder with no remote repository behind it, as they used to with a Repo. What happens?
Requiring a remote URL is one of the few real differences between Git folders and legacy Repos.
“Git folders require a remote repository URL. Repos didn't have this requirement.”Source: docs.databricks.com
2.Connecting your Git provider
Before you can push anything, Databricks needs a Git credential for your provider. You can clone a public repository without one. Writing to any repository, or working with a private one, needs a credential with Write permissions.
Git folders support cloud providers: GitHub (including GitHub Advanced Enterprise and GitHub Enterprise Cloud), Atlassian Bitbucket Cloud, GitLab and GitLab Enterprise Edition, Microsoft Azure DevOps (Azure Repos), and AWS CodeCommit. They also support on-premises providers: GitHub Enterprise Server, Bitbucket Server and Data Center, and GitLab Self-Managed. Some providers support OAuth account linking, and others need a personal access token (PAT). For hosted GitHub, Databricks recommends the Databricks GitHub app over PATs. GitHub Enterprise Server can't link an account, so it needs a PAT.
Git folders connect only over HTTPS, not SSH. They also don't support GPG signing of commits.
Checkpoint 2 of 5· Put it in order
Put the steps for adding a PAT-based Git credential in the Databricks UI in order.
- 1.Click Save
- 2.Enter your email and paste your PAT in the Token field
- 3.Select your Git provider from the drop-down menu
- 4.Click Add Git credential
- 5.Click Linked accounts
- 6.Click your username in the top bar and select Settings
You add credentials under your user Settings > Linked accounts. There you pick the provider and then supply the email and token.
“Click your username in the top bar of the Databricks workspace and select Settings.”Source: docs.databricks.com
The email on that credential matters more than it looks. It controls how your commits appear in the provider. The email becomes the author email (GIT_AUTHOR_EMAIL and GIT_COMMITTER_EMAIL), and the username becomes the committer name (GIT_AUTHOR_NAME and GIT_COMMITTER_NAME). If you leave the email out, Databricks uses your Git username as the email, and the provider may not link the commits to your profile. Each user can store up to 10 Git credentials, with one default per provider. A specific Git folder can be pointed at a non-default credential from its Git settings tab.
3.Cloning a repository into a Git folder
Cloning creates the Git folder. Two preconditions apply: your workspace has Git credentials configured, and you have CAN MANAGE permission on the parent folder. In the UI, select Workspace in the sidebar, browse to the parent folder, and click Create > Git folder. The dialog asks for the Git repository URL, the Git provider, the Git folder name, and whether to use sparse checkout mode, which clones only a subset of directories. Click Create Git folder, and the remote contents are cloned into your workspace.
A Public Preview feature can also give a Git folder Git CLI access, which lets you run standard Git commands. When your workspace is eligible, a folder created from the UI gets it automatically. If the workspace isn't eligible, Databricks creates a standard Git folder, and the Git dialog still works. You can also clone from the web terminal, and a folder cloned there always has Git CLI access:
git clone <remote-url>Checkpoint 3 of 5· Check yourself
A user has valid Git credentials but gets a permission error when they try to create a Git folder inside a shared project folder. What is the most likely cause?
Creating a Git folder requires CAN MANAGE on its parent folder. HTTPS is the only supported protocol, and Git folders can live at any depth under your user folder.
“You must have CAN MANAGE permission on the parent folder where you want to create the Git folder.”Source: docs.databricks.com
Sources5
4.The Git dialog: where daily work happens
Day-to-day UI work goes through one full-screen Git dialog, and there are two ways to open it. From a notebook, click the button next to the notebook name; it shows the current Git branch. From the Git folders browser, click Git beside the repo name. Either way, you get the same controls.
| Control | What it does |
|---|---|
| Branch selector | Shows your current working branch and lets you select another |
| Create Branch | Creates a new branch from a base branch you choose |
| Changed files | Lists file assets and subfolders checked into the current branch |
| History | Shows the current branch history |
| Pull | Pulls content from the remote Git repository |
| Commit message and description | A commit message and an optional expanded description |
| Commit and push | Commits to the working branch and pushes the updated branch to the remote |
| Kebab menu | Other branch operations: hard reset, merge, rebase |
Checkpoint 4 of 5· Match them up
Match each Git dialog control to what you would use it for.
Tap a term, then the definition that fits it.
The dialog keeps the daily operations (switch, create, pull, commit and push) visible and puts the less frequent branch operations in the kebab menu.
“Click the kebab menu to choose from additional Git branch operations, such as a hard reset, merge, or rebase.”Source: docs.databricks.com
Checkpoint 5 of 5· Check yourself
You have finished work on a feature branch in a Git folder and want it merged into the main branch. Which statement is correct?
Git folders support branch merging, and a pull request merged through the provider is the alternative. Deleting a branch is the part you must do in your Git provider.
“Git folders support branch merging. You can also create a pull request and merge through your Git provider.”Source: docs.databricks.com
Exam traps
Each one states something that sounds right. Open it to see what is actually true.
1.Git folders can connect to the remote over SSH, the same way a local Git client can.Why is that wrong?
Git folders use HTTPS only and authenticate with a stored credential (OAuth link or PAT). They also don't support GPG-signed commits.
Covered in Connecting your Git provider
2.Moving from Repos to Git folders changed how branching, committing, and pushing work.Why is that wrong?
The Git operations themselves are the same. What changed is the name, the folder placement rules, the required remote repository URL, and the asset types a folder can hold.
Covered in From Repos to Git folders
Sources
Every claim above is drawn from one of these pages, quoted as it was written on the date shown.
- 1.
“Repos are now called Git folders.”
↩︎ From Repos to Git folders“Repos could only exist at a fixed level like /Workspace/Repos/<user email>/<Repo name>.”
↩︎ From Repos to Git folders“Git folders can exist at any depth under /Workspace/Users/<user-email>/, and you can have multiple Git folders.”
↩︎ From Repos to Git folders“Most Git operations require the CAN MANAGE permission.”
↩︎ From Repos to Git folders“The core functionality hasn't changed, only the terminology.”
↩︎ Exam trap 2“The core functionality hasn't changed, only the terminology.”
↩︎ Prediction“Git folders require a remote repository URL. Repos didn't have this requirement.”
↩︎ Checkpoint - 2.
“Create and manage branches for development work, including merging, rebasing, and resolving conflicts.”
↩︎ From Repos to Git folders - 3.https://docs.databricks.com/aws/en/repos/repos-setupOfficial docs
“To modify a public remote repository or to work with private repositories, configure Git credentials with Write permissions.”
↩︎ Connecting your Git provider“Username: Becomes the committer name (GIT_AUTHOR_NAME and GIT_COMMITTER_NAME)”
↩︎ Connecting your Git provider“If you don't specify an email address, Databricks uses your Git username as the email.”
↩︎ Connecting your Git provider“Each user can have a maximum of 10 Git credentials.”
↩︎ Connecting your Git provider“Click your username in the top bar of the Databricks workspace and select Settings.”
↩︎ Checkpoint - 4.
“For hosted GitHub accounts, Databricks recommends using the Databricks GitHub app for authentication instead of personal access tokens (PATs).”
↩︎ Connecting your Git provider - 5.
“Your workspace must have Git credentials configured.”
↩︎ Cloning a repository into a Git folder“Git folders that you clone from the web terminal always have Git CLI access.”
↩︎ Cloning a repository into a Git folder“From the Databricks Git folders browser, click Git beside the repo name.”
↩︎ The Git dialog: where daily work happens“From a notebook, click the button next to the notebook name that identifies the current Git branch.”
↩︎ The Git dialog: where daily work happens“You must have CAN MANAGE permission on the parent folder where you want to create the Git folder.”
↩︎ Checkpoint“Click the kebab menu to choose from additional Git branch operations, such as a hard reset, merge, or rebase.”
↩︎ Checkpoint - 6.https://docs.databricks.com/aws/en/repos/limitsOfficial docs
“To delete a branch, you must work in your Git provider.”
↩︎ The Git dialog: where daily work happens“Git folders support only HTTPS, not SSH.”
↩︎ Exam trap 1“Git folders support branch merging. You can also create a pull request and merge through your Git provider.”
↩︎ Checkpoint
Also cited
- https://docs.databricks.com/aws/en/reposOfficial docs
“Databricks Git folders is a visual Git client and API that integrates Git repositories within your workspace.”
↩︎ Key concept