Git Cheat Sheet

Git lets you manage changes made to files and directories in a project. It allows you to keep track of what you did, undo any changes you decide you don’t want and collaborate at scale with others. This cheat sheet will give you a download on all things Git.

Version control tools like Git are essential for working with data science in the real world. Git helps data scientists track their work, collaborate with others easily, experiment with changes to their existing work in a safe environment, and more. This cheat sheet will give you a download on all things Git.

What is Version Control?

Version control systems are tools that manage changes made to files and directories in a project. They allow you to keep track of what you did when undoing any changes you decide you don’t want and collaborate at scale with others. This cheat sheet focuses on one of the most popular ones, Git.

Key Definitions

Throughout this cheat sheet, you’ll find git-specific terms and jargon being used. Here’s a run-down of all the terms you may encounter

Basic definitions

• Local repo or repository: A local directory containing code and files for the project
• Remote repository: An online version of the local repository hosted on services like GitHub, GitLab, and BitBucket
• Cloning: The act of making a clone or copy of a repository in a new directory
• Commit: A snapshot of the project you can come back to
• Branch: A copy of the project used for working in an isolated environment without affecting the main project
• Git merge: The process of combining two branches together

More advanced definitions

• .gitignore file: A file that lists other files you want git not to track (e.g. large data folders, private info, and any local files that shouldn’t be seen by the public.)
• Staging area: a cache that holds changes you want to commit next.
• Git stash: another type of cache that holds unwanted changes you may want to come back later
• Commit ID or hash: a unique identifier for each commit, used for switching to different save points.
• HEAD (always capitalized letters): a reference name for the latest commit, to save you having to type Commit IDs. HEAD~n syntax is used to refer to older commits (e.g. HEAD~2 refers to the second-to-last commit).

Installing Git

On OS-X — Using an Installer

• Download the installer for Mac
• Follow the prompts

On OS-X — Using Homebrew

brew install git

On Linux

sudo apt-get install git

On Windows

• Download the latest Git For Windows installer
• Follow the prompts

Check if installation is successful on any platform

git --version

Setting Up Git

If you are working in a team on a single repo, it is important for others to know who made certain changes to the code. So, Git allows you to set user credentials such as name, email, etc…

Set your basic information

Configure your email

git config user.email [your.email@domain.com]

Configure your name

git config user.name [your-name]

Important tags to determine the scope of configurations

Git lets you use tags to determine the scope of the information you’re using during setup

Local directory, single project (this is the default tag)

git config --local user.email “my_email@example.com”

All git projects under the current user

git config --global user.email “my_email@example.com”

For all users on the current machine

git config --system user.email “my_email@example.com”

Other useful configuration commands

List all key-value configurations

git config --list

Get the value of a single key

git config --get <key>

Setting aliases for common commands

If you find yourself using a command frequently, git lets you set an alias for that command to surface it more quickly

Create an alias named gc for the “git commit” command

git config --global alias.gc commit

gc -m “New commit”

Create an alias named ga for the “git add” command

git config --global alias.ga add

What is a Branch?

Branches are special “copies” of the code base which allow you to work on different parts of a project and new features in an isolated environment. Changes made to the files in a branch won’t affect the “main branch” which is the main project development channel.

Git Basics

What is a repository?

A repository or a repo is any location that stores code and the necessary files that allow it to run without errors. A repo can be both local and remote. A local repo is typically a directory on your machine while a remote repo is hosted on servers like GitHub

Creating local repositories

Clone a repository from remote hosts (GitHub, GitLab, DagsHub, etc.)

git clone <remote_repo_url>

Initialize git tracking inside the current directory

git init

Create a git-tracked repository inside a new directory

git init [dir_name]

Clone only a specific branch

git clone -branch <branch_name> <repo_url>

Cloning into a specified directory

git clone <repo_url> <dir_name>

A note on cloning

There are two primary methods of cloning a repository – HTTPS syntax and SSH syntax. While SSH cloning is generally considered a bit more secure because you have to use an SSH key for authentication, HTTPS cloning is much simpler and the recommended cloning option by GitHub.

HTTPS

git clone https://github.com/your_username/repo_name.git

SSH

git clone git@github.com:user_name/repo_name.git

Managing remote repositories

List remote repos

git remote

Create a new connection called <remote> to a remote repository on servers like GitHub, GitLab, DagsHub, etc.

git remote add <remote> <url_to_remote>

Remove a connection to a remote repo called <remote>

git remote rm <remote>

Rename a remote connection

git remote rename <old_name> <new_name>

Working with Files

Adding and removing files

Add a file or directory to git for tracking

git add <filename_or_dir>

Add all untracked and tracked files inside the current directory to git

git add .

Remove a file from a working directory or staging area

git rm <<filename_or_dir>

Saving and working with changes

See changes in the local repository

git status

Saving a snapshot of the staged changes with a custom message

git commit -m “[Commit message]”

Staging changes in all tracked files and committing with a message

git add -am “[Commit message]”

Editing the message of the latest commit

git commit --amend -m “[New commit message]”

Saving staged and unstaged changes to stash for later use (see below for the explanation of a stash)

git stash

Stashing staged, unstaged and untracked files as well

git stash -u

Stashing everything (including ignored files)

git stash --all

Reapply previously stashed changes and empty the stash

git stash pop

Reapply previously stashed changes and keep the stash

git stash apply

Dropping changes in the stash

git stash drop

Show uncommitted changes since the last commit

git diff

Show the differences between two commits (should provide the commit IDs)

git diff <id_1> <id_2>

A note on stashes

Git stash allows you to temporarily save edits you’ve made to your working copy so you can return to your work later. Stashing is especially useful when you are not yet ready to commit changes you’ve done, but would like to revisit them at a later time.

Branches

List all branches

git branch

git branch --list

git branch -a  (shows remote branches as well)

Create a new local branch named new_branch without checking out that branch

git branch <new_branch>

Switch into an existing branch named <branch>

git checkout <branch>

Create a new local branch and switch into it

git checkout -b <new_branch>

Safe delete a local branch (prevents deleting unmerged changes)

git branch -d <branch>

Force delete a local branch (whether merged or unmerged)

git branch -D <branch>

Rename the current branch to <new_name>

git branch -m <new_name>

Push a copy of local branch named branch to the remote repo

git push <remote_repo> branch~

Delete a remote branch named branch (-d tag only works locally)

git push <remote_repo> :branch

git push <remote_repo> --delete branch

Merging a branch into the main branch

git checkout main

git merge <other_branch>

Merging a branch and creating a commit message

git merge --no-ff <other_branch>

Compare the differences between two branches

git diff <branch_1> <branch_2>

Compare a single <file> between two branches

git diff <branch_1> <branch_2> <file>

Pulling changes

 

Download all commits and branches from the <remote> without applying them on the local repo

git fetch <remote>

Only download the specified <branch> from the <remote>

git fetch <remote> <branch>

Merge the fetched changes if accepted

git merge <remote>/<branch>

A more aggressive version of fetch which calls fetch and merge simultaneously

git pull <remote>

Logging and reviewing work

 

List all commits with their author, commit ID, date, and message

git log

List one commit per line (-n tag can be used to limit the number of commits displayed (e.g. -5))

git log --oneline [-n]

Log all commits with diff information

git log --stat

Log commits after some date (A sample value can be 4th of October, 2020 – “2020-10-04” or keywords such as “yesterday”, “last month”, etc. )

git log --oneline --after=”YYYY-MM-DD”

Log commits before some date (Both –after and –before tags can be used for date ranges)

git log --oneline --before=”last year”

Reversing changes

Checking out (switching to) older commits

git checkout HEAD~3

Checks out the third-to-last commit.

git checkout <commit_id>

Undo the latest commit but leave the working directory unchanged

git reset HEAD~1

You can undo as many commits as you want by changing the number after the tilde.

Discard all changes of the latest commit (no easy recovery)

git reset --hard HEAD~1

Instead of HEAD~n, you can provide commit hash as well. Changes after that commit will be destroyed.

Undo a single given commit, without modifying commits that come after it (a safe reset)

git revert [commit_id]

May result in revert conflicts

Download Git Cheat Sheet