Git – How to clone a single branch in Git

branchgitgit-clone

I have a local Git repository called 'skeleton' that I use for storing project skeletons. It has a few branches, for different kinds of projects:

casey@agave [~/Projects/skeleton] git branch
* master
  rails
  c
  c++

If I want to check out the master branch for a new project, I can do

casey@agave [~/Projects] git clone skeleton new
Initialized empty Git repository in /Users/casey/Projects/new/.git/

and everything is how I want it. Specifically, the new master branch points to the skeleton master branch, and I can push and pull to move around changes to the basic project setup.

What doesn't work, however, is if I want to clone another branch. I can't get it so that I only pull the branch I want, for instance the rails branch, and then the new repository has a master branch that pushes to and pulls from the skeleton repository's rails branch, by default.

Is there a good way to go about doing this? Or, maybe this isn't the way that Git wants me to structure things, and I'm certainly open to that. Perhaps I should have multiple repositories, with the Ruby on Rails skeleton repository tracking the master skeleton repository? And any individual project cloning the Ruby on Rails skeleton repository.

Best Answer

Note: the git1.7.10 (April 2012) actually allows you to clone only one branch:

# clone only the remote primary HEAD (default: origin/master)
git clone <url> --single-branch

# as in:
git clone <url> --branch <branch> --single-branch [<folder>]

(<url> is the URL if the remote repository, and does not reference itself the branch cloned)

You can see it in t5500-fetch-pack.sh:

test_expect_success 'single branch clone' '
  git clone --single-branch "file://$(pwd)/." singlebranch
'

Tobu comments that:

This is implicit when doing a shallow clone.
This makes git clone --depth 1 the easiest way to save bandwidth.

And since Git 1.9.0 (February 2014), shallow clones support data transfer (push/pull), so that option is even more useful now.
See more at "Is git clone --depth 1 (shallow clone) more useful than it makes out?".


"Undoing" a shallow clone is detailed at "Convert shallow clone to full clone" (git 1.8.3+)

# unshallow the current branch
git fetch --unshallow

# for getting back all the branches (see Peter Cordes' comment)
git config remote.origin.fetch refs/heads/*:refs/remotes/origin/*
git fetch --unshallow

As Chris comments:

the magic line for getting missing branches to reverse --single-branch is (git v2.1.4):

git config remote.origin.fetch +refs/heads/*:refs/remotes/origin/*
git fetch --unshallow  

With Git 2.26 (Q1 2020), "git clone --recurse-submodules --single-branch" now uses the same single-branch option when cloning the submodules.

See commit 132f600, commit 4731957 (21 Feb 2020) by Emily Shaffer (nasamuffin).
(Merged by Junio C Hamano -- gitster -- in commit b22db26, 05 Mar 2020)

clone: pass --single-branch during --recurse-submodules

Signed-off-by: Emily Shaffer
Acked-by: Jeff King

Previously, performing "git clone --recurse-submodules --single-branch" resulted in submodules cloning all branches even though the superproject cloned only one branch.

Pipe --single-branch through the submodule helper framework to make it to 'clone' later on.