要更新现有的远程跟踪特定的分支，只使用:

git remote set-branches <remote-name> <branch-name>

从git帮助远程:

set-branches
    Changes the list of branches tracked by the named remote. This can be used to track a subset of the available remote branches
    after the initial setup for a remote.

    The named branches will be interpreted as if specified with the -t option on the git remote add command line.

    With --add, instead of replacing the list of currently tracked branches, adds to that list.

让我用我的两便士来曲解马德森先生的回答:

Git获取<remote_name_or_url> <branch_name> git checkout FETCH_HEAD -B <branch_name>

与马德森的提议相比，这两条线路的主要优势在于，即使分支机构已经在本地存在，它也能工作。

一种方法是:

在.git/config中获取远程回购应该设置为获取任何分支:

   [remote "origin"]
            fetch = +refs/heads/*:refs/remotes/origin/*

获取远程分支:

git fetch origin branch-name

创建一个本地分支'branch-name'，用于从原点跟踪远程分支'branch-name'。

git checkout -b branch-name origin/branch-name

列出所有分支

git branch -a

这是最简单的方法

  git fetch origin <branch> && git checkout <branch>

示例:我想从原点获取uat分支，并切换到此作为当前工作分支。

   git fetch origin uat && git checkout uat

答案取决于你想要完成什么。

If it is a one time deal from a different repository and you don't need a reference (e.g. to merge GitHub Pull Requests, where <remote_url> is like https://github.com/USER/REPO.git), then you can use: git checkout -b <local_branch> <local_branch_to merge_into> git pull <remote_url> <remote_branch> If you want to update and track the branch you have to set first the remote and there are 4 alternatives: If you are cloning a new repository (e.g. working only on it) git clone --single-branch --branch remote_branch remote_url If you are adding a new remote to your working directory # multiple -t options are allowed git remote add -t <remote_branch> <remote_repo> <remote_url> If you are adding the branch restriction to an existing remote in your working directory # with --add it will add the branch instead of setting it # you can add multiple branches with multiple --add lines # wildcards are allowed, # e.g. branch_v\* matching branch_v1, branch_v2, ... git remote set-branches [--add] <remote_repo> <remote_branch> You could also skip the restrictions because clone by default fetches only the main branch and remote add does not fetch branches. Buth then you'll have to mention the remote branch all the remote branch all the time you fetch the remote_repo. git remote add <remote_repo> <remote_url> After setting the remote you can fetch the remote branch, checkout and pull: # If you set only one <remote_branch> in the restrictions above (i.e no option 4), # then you can omit it and still only <remote_branch> will be fetched git fetch <remote_repo> [<remote_branch>] # without -b the local branch name is guessed to be the same as the remote one git checkout --track [-b <local_branch>] <remote_repo>/<remote_branch>

检查远程和已经或将要获取的分支的最佳命令是git remote show <remote_repo>。它会打印“Remote branch:”下的分支列表，还会告诉你它们是否被获取，以及是否被跟踪。

你也可以通过使用git branch -r列出已知的远程分支来检查远程分支的限制，如果你有很多远程，可以结合grep，或者通过检查git配置文件.git/config中的远程详细信息。它将包含如下部分:

[remote "<remote_repo>"]
    url = <remote_url>
    fetch = +refs/heads/<remote_branch>:refs/remotes/<remote_repo>/<remote_branch>

编辑配置文件将改变限制，但我同意@alexk的意见，这不是一个好主意。

NOTE: If a branch is not in the list of branches of a remote (visible in git remote show or the config file), then you will not be able to have a reference to it, git will save it to the temporary FETCH_HEAD and you will not be able to track it or to use it directly in git checkout. This is the problem that brought me to this thread (the opposite of the one in the question): I cloned a repo with GitHub client gh repo clone USER/REPO and it added automatically "upstream", the repository forked from, restricted only to the branch "master". I was unable to checkout other branches and getting errors like "fatal: '<remote_repo>/<remote_branch>' is not a commit and a branch '<local_branch>' cannot be created from it". I fixed it with: git remote set-branches <remote_repo> \*.

在我的例子中，我想在不创建新的远程的情况下获取一个分支，所以这是有效的:

Git获取<远程url> <远程分支名称>:<本地分支名称>