问题

How can I stash only one of multiple changed files on my branch?

回答1:

Update: the following answer is for git before git 2.13. For git 2.13 and over, check out another answer further down.

Warning

As noted in the comments, this puts everything into the stash, both staged and unstaged. The --keep-index just leaves the index alone after the stash is done. This can cause merge conflicts when you later pop the stash.

This will stash everything that you haven't previously added. Just git add the things you want to keep, then run it.

git stash --keep-index

For example, if you want to split an old commit into more than one changeset, you can use this procedure:

git rebase -i <last good commit>
Mark some changes as edit.
git reset HEAD^
git add <files you want to keep in this change>
git stash --keep-index
Fix things up as necessary. Don't forget to git add any changes.
git commit
git stash pop
Repeat, from #5, as necessary.
git rebase --continue

回答2:

You can also use git stash save -p "my commit message". This way you can select which hunks should be added to stash, whole files can be selected as well.

You'll be prompted with a few actions for each hunk:

   y - stash this hunk
   n - do not stash this hunk
   q - quit; do not stash this hunk or any of the remaining ones
   a - stash this hunk and all later hunks in the file
   d - do not stash this hunk or any of the later hunks in the file
   g - select a hunk to go to
   / - search for a hunk matching the given regex
   j - leave this hunk undecided, see next undecided hunk
   J - leave this hunk undecided, see next hunk
   k - leave this hunk undecided, see previous undecided hunk
   K - leave this hunk undecided, see previous hunk
   s - split the current hunk into smaller hunks
   e - manually edit the current hunk
   ? - print help

回答3:

Since git is fundamentally about managing a all repository content and index (and not one or several files), git stash deals, not surprisingly, ~~with the all working directory~~.

Actually, since Git 2.13 (Q2 2017), you can stash individual files, with git stash push:

git stash push [--] [<pathspec>...]

When pathspec is given to 'git stash push', the new stash records the modified states only for the files that match the pathspec

See "Stash changes to specific files" for more.

The test case is self-explanatory:

test_expect_success 'stash with multiple pathspec arguments' '
    >foo &&
    >bar &&
    >extra &&
    git add foo bar extra &&

    git stash push -- foo bar &&   

    test_path_is_missing bar &&
    test_path_is_missing foo &&
    test_path_is_file extra &&

    git stash pop &&
    test_path_is_file foo &&
    test_path_is_file bar &&
    test_path_is_file extra

The original answer (below, June 2010) was about manually selecting what you want to stash.

Casebash comments:

This (the stash --patch original solution) is nice, but often I've modified a lot of files so using patch is annoying

bukzor's answer (upvoted, November 2011) suggests a more practical solution, based on
git add + git stash --keep-index.
Go see and upvote his answer, which should be the official one (instead of mine).

About that option, chhh points out an alternative workflow in the comments:

you should "git reset --soft" after such a stash to get your clear staging back:
In order to get to the original state - which is a clear staging area and with only some select un-staged modifications, one could softly reset the index to get (without committing anything like you - bukzor - did).

(Original answer June 2010: manual stash)

Yet, git stash save --patch could allows you to achieve the partial stashing you are after:

With --patch, you can interactively select hunks from in the diff between HEAD and the working tree to be stashed.
The stash entry is constructed such that its index state is the same as the index state of your repository, and its worktree contains only the changes you selected interactively. The selected changes are then rolled back from your worktree.

However that will save the full index (which may not be what you want since it might include other files already indexed), and a partial worktree (which could look like the one you want to stash).

git stash --patch --no-keep-index

might be a better fit.

If --patch doesn't work, a manual process might:

For one or several files, an intermediate solution would be to:

copy them outside the Git repo
(Actually, eleotlecram proposes an interesting alternative)
git stash
copy them back
git stash # this time, only the files you want are stashed
git stash pop stash@{1} # re-apply all your files modifications
git checkout -- afile # reset the file to the HEAD content, before any local modifications

At the end of that rather cumbersome process, you will have only one or several files stashed.

回答4:

When git stash -p (or git add -p with stash --keep-index) would be too cumbersome, I found it easier to use diff, checkout and apply:

To "stash" a particular file/dir only:

git diff path/to/dir > stashed.diff
git checkout path/to/dir

Then afterwards

git apply stashed.diff

回答5:

Use git stash push, like this:

git stash push [--] [<pathspec>...]

For example:

git stash push -- my/file.sh

This is available since Git 2.13, released in spring 2017.

回答6:

Let's say you have 3 files

a.rb
b.rb
c.rb

and you want to stash only b.rb and c.rb but not a.rb

you can do something like this

# commit the files temporarily you don't want to stash
git add a.rb
git commit -m "temp" 

# then stash the other files
git stash save "stash message"

# then undo the previous temp commit
git reset --soft HEAD^
git reset

And you are done! HTH.

回答7:

Another way to do this:

# Save everything
git stash 

# Re-apply everything, but keep the stash
git stash apply

git checkout <"files you don't want in your stash">

# Save only the things you wanted saved
git stash

# Re-apply the original state and drop it from your stash
git stash apply stash@{1}
git stash drop stash@{1}

git checkout <"files you put in your stash">

I came up with this after I (once again) came to this page and didn't like the first two answers (the first answer just doesn't answer the question and I didn't quite like working with the -p interactive mode).

The idea is the same as what @VonC suggested using files outside the repository, you save the changes you want somewhere, remove the changes you don't want in your stash, and then re-apply the changes you moved out of the way. However, I used the git stash as the "somewhere" (and as a result, there's one extra step at the end: removing the cahnges you put in the stash, because you moved these out of the way as well).

回答8:

Update (2/14/2015) - I've rewritten the script a bit, to better handle the case of conflicts, which should now be presented as unmerged conflicts rather than .rej files.

I often find it more intuitive to do the inverse of @bukzor's approach. That is, to stage some changes, and then stash only those staged changes.

Unfortunately, git doesn't offer a git stash --only-index or similar, so I whipped up a script to do this.

#!/bin/sh

# first, go to the root of the git repo
cd `git rev-parse --show-toplevel`

# create a commit with only the stuff in staging
INDEXTREE=`git write-tree`
INDEXCOMMIT=`echo "" | git commit-tree $INDEXTREE -p HEAD`

# create a child commit with the changes in the working tree
git add -A
WORKINGTREE=`git write-tree`
WORKINGCOMMIT=`echo "" | git commit-tree $WORKINGTREE -p $INDEXCOMMIT`

# get back to a clean state with no changes, staged or otherwise
git reset -q --hard

# Cherry-pick the index changes back to the index, and stash.
# This cherry-pick is guaranteed to succeed
git cherry-pick -n $INDEXCOMMIT
git stash

# Now cherry-pick the working tree changes. This cherry-pick may fail
# due to conflicts
git cherry-pick -n $WORKINGCOMMIT

CONFLICTS=`git ls-files -u`
if test -z "$CONFLICTS"; then
    # If there are no conflicts, it's safe to reset, so that
    # any previously unstaged changes remain unstaged
    #
    # However, if there are conflicts, then we don't want to reset the files
    # and lose the merge/conflict info.
    git reset -q
fi

You can save the above script as git-stash-index somewhere on your path, and can then invoke it as git stash-index

# <hack hack hack>
git add <files that you want to stash>
git stash-index

Now the stash contains a new entry that only contains the changes you had staged, and your working tree still contains any unstaged changes.

In some cases, the working tree changes may depend on the index changes, so when you stash the index changes, the working tree changes have a conflict. In this case, you'll get the usual unmerged conflicts that you can resolve with git merge/git mergetool/etc.

回答9:

Since creating branches in Git is trivial you could just create a temporary branch and check the individual files into it.

回答10:

If you do not want to specify a message with your stashed changes, pass the filename after a double-dash.

$ git stash -- filename.ext

If it's an untracked/new file, you will have to stage it first.

This method works in git versions 2.13+

回答11:

Just in case you actually mean discard changes whenever you use git stash (and don't really use git stash to stash it temporarily), in that case you can use

git checkout -- <file>

[NOTE]

That git stash is just a quicker and simple alternative to branching and doing stuff.

回答12:

Save the following code to a file, for example, named stash. Usage is stash <filename_regex>. The argument is the regular expression for the full path of the file. For example, to stash a/b/c.txt, stash a/b/c.txt or stash .*/c.txt, etc.

$ chmod +x stash
$ stash .*.xml
$ stash xyz.xml

Code to copy into the file:

#! /usr/bin/expect --
log_user 0
set filename_regexp [lindex $argv 0]

spawn git stash -p

for {} 1 {} {
  expect {
    -re "diff --git a/($filename_regexp) " {
      set filename $expect_out(1,string)
    }
    "diff --git a/" {
      set filename ""
    }
    "Stash this hunk " {
      if {$filename == ""} {
        send "n\n"
      } else {
        send "a\n"
        send_user "$filename\n"
      }
    }
    "Stash deletion " {
      send "n\n"
    }
    eof {
      exit
    }
  }
}

回答13:

You can simple do this:

git stash push "filename"

or with an optional message

git stash push -m "Some message" "filename"

回答14:

The problem with VonC's `intermediate' solution of copying files to outside the Git repo is that you lose path information, which makes copying a bunch of files back later on somewhat of a hassle.

A find it easier to use tar (similar tools will probably do) instead of copy:

tar cvf /tmp/stash.tar path/to/some/file path/to/some/other/file (... etc.)
git checkout path/to/some/file path/to/some/other/file
git stash
tar xvf /tmp/stash.tar
etc. (see VonC's `intermediate' suggestion)

回答15:

Sometimes I've made an unrelated change on my branch before I've committed it, and I want to move it to another branch and commit it separately (like master). I do this:

git stash
git checkout master
git stash pop
git add <files that you want to commit>
git commit -m 'Minor feature'
git stash
git checkout topic1
git stash pop
...<resume work>...

Note the first stash & stash pop can be eliminated, you can carry all of your changes over to the master branch when you checkout, but only if there are no conflicts. Also if you are creating a new branch for the partial changes you will need the stash.

You can simplify it assuming no conflicts and no new branch:

git checkout master
git add <files that you want to commit>
git commit -m 'Minor feature'
git checkout topic1
...<resume work>...

Stash not even needed...

回答16:

This can be done easily in 3 steps using SourceTree.

Temporarily commit everything you don't want stashed.
Git add everything else, then stash it.
Pop your temporary commit by running git reset, targetting the commit before your temporary one.

This can all be done in a matter of seconds in SourceTree, where you can just click on the files (or even individual lines) you want to add. Once added, just commit them to a temporary commit. Next, click the checkbox to add all changes, then click stash to stash everything. With the stashed changes out of the way, glance over at your commit list and note the hash for the commit before your temporary commit, then run 'git reset hash_b4_temp_commit', which is basically like "popping" the commit by resetting your branch to the commit right before it. Now, you're left with just the stuff you didn't want stashed.

回答17:

I would use git stash save --patch. I don't find the interactivity to be annoying because there are options during it to apply the desired operation to entire files.

回答18:

Every answer here is so complicated...

What about this to "stash":

git diff /dir/to/file/file_to_stash > /tmp/stash.patch
git checkout -- /dir/to/file/file_to_stash

This to pop the file change back:

git apply /tmp/stash.patch

Exact same behavior as stashing one file and popping it back in.

回答19:

I've reviewed answers and comments for this and a number of similar threads. Be aware that none of the following commands are correct for the purpose of being able to stash any specific tracked/untracked files:

git stash -p (--patch): select hunks manually, excluding untracked files
git stash -k (--keep-index): stash all tracked/untracked files and keep them in the working directory
git stash -u (--include-untracked): stash all tracked/untracked files
git stash -p (--patch) -u (--include-untracked): invalid command

Currently, the most reasonable method to be able to stash any specific tracked/untracked files is to:

Temporarily commit the files you don't want to stash
Add and stash
Pop the temporary commit

I wrote a simple script for this procedure in an answer to another question, and there are steps for performing the procedure in SourceTree here.

回答20:

Solution

Local changes:

file_A (modified) not staged
file_B (modified) not staged
file_C (modified) not staged

To create a stash "my_stash" with only the changes on file_C:

1. git add file_C
2. git stash save --keep-index temp_stash
3. git stash save my_stash
4. git stash pop stash@#{1}

Done.

Explanation

add file_C to the staging area
create a temporary stash named "temp_stash" and keep the changes on file_C
create the wanted stash ("my_stash") with only the changes on file_C
apply the changes in "temp_stash" (file_A and file_B) on your local code and delete the stash

You can use git status between the steps to see what is going on.

回答21:

In this situation I git add -p (interactive), git commit -m blah and then stash what's left if necessary.

回答22:

When you try to switch between two branches, this situation occurs.

Try to add the files using "git add filepath".

Later execute this line

git stash --keep-index

回答23:

To stash a single file use git stash --patch [file].

This is going to prompt: Stash this hunk [y,n,q,a,d,j,J,g,/,e,?]? ?. Just type a (stash this hunk and all later hunks in the file) and you're fine.

回答24:

git add .                           //stage all the files
git reset <pathToFileWillBeStashed> //unstage file which will be stashed
git stash                           //stash the file(s)
git reset .                         // unstage all staged files
git stash pop                       // unstash file(s)

回答25:

Similar situation. Did commit and realized it's not ok.

git commit -a -m "message"
git log -p

Based on the answers this helped me.

# revert to previous state, keeping the files changed
git reset HEAD~
#make sure it's ok
git diff
git status
#revert the file we don't want to be within the commit
git checkout specs/nagios/nagios.spec
#make sure it's ok
git status
git diff
#now go ahead with commit
git commit -a -m "same|new message"
#eventually push tu remote
git push

回答26:

I don't know how to do it on command line, only using SourceTree. Lets say you have changed file A, and have two change hunks in file B. If you want to stash only the second hunk in file B and leave everything else untouched, do this:

Stage everything
Perform changes to your working copy that undo all the changes in file A. (e.g. launch external diff tool and make files match.)
Make file B look as if only second change is applied to it. (e.g. launch external diff tool and undo first change.)
Create a stash using "Keep staged changes".
Unstage everything
Done!

回答27:

One complicated way would be to first commit everything:

git add -u
git commit // creates commit with sha-1 A

Reset back to the original commit but checkout the_one_file from the new commit:

git reset --hard HEAD^
git checkout A path/to/the_one_file

Now you can stash the_one_file:

git stash

Cleanup by saving the committed content in your file system while resetting back to the original commit:

git reset --hard A
git reset --soft HEAD^

Yeah, somewhat awkward...

回答28:

I found no answer to be what I needed and that is as easy as:

git add -A
git reset HEAD fileThatYouWantToStash
git commit -m "committing all but one file"
git stash

This stashes exactly one file.

回答29:

Quick Answer

For revert a specific changed file in git, you can do the following line:

git checkout <branch-name> -- <file-path>

Here is an actual example:

git checkout master -- battery_monitoring/msg_passing.py

回答30:

If you want to stash some changed files, simply just

Add the files you don't want to stash, in Stage, Then execute git stash save --keep-index

It will stash all unstaged changed files

来源：https://stackoverflow.com/questions/3040833/stash-only-one-file-out-of-multiple-files-that-have-changed-with-git

标签

git

git-stash

Stash only one file out of multiple files that have changed with Git?

问题