Remove all commits but keep all file changes

Question

I have 10 commits on my branch with having 15 files changed. Now I want to keep only file changes but not commits.

Means Now I want 15 file changes in one commit.

git reset --hard commit_id
git push --force

The above command will not help us.

Edit:

I want to move my branch to one specific commit, with maintaining all file changes which happened in current commits. Then Want to add all unstash changes in one commit.

Can it be possible?

Check the duplicate link for a number of ways to _squash_ your 10 commits into a single commit. — Tim Biegeleisen, Jun 20 '19 at 13:48
@TimBiegeleisen: I don't think the above link will help me, can you share the working code for it? — Jitesh Mohite, Jun 20 '19 at 14:07
`git reset --soft HEAD~10; git commit -m "there, 10 squashed commits"` — eftshift0, Jun 20 '19 at 14:08
Either do an interactive rebase with a squash, or try doing a soft reset. — Tim Biegeleisen, Jun 20 '19 at 14:10
@TimBiegeleisen: I am not getting my question answer dere, also I don't want to mention specific commit numbers. How it will be possible then? — Jitesh Mohite, Jun 20 '19 at 14:15
Try the root commit if you want to squash everything (from a specific branch that is): `git reset --soft $(git rev-list --max-parents=0 HEAD)` — evolutionxbox, Jun 20 '19 at 16:29
Please edit your question and tell us what the actual problem is. _Which_ commits do you want to squash, and where are they located? — Tim Biegeleisen, Jun 21 '19 at 11:07
This question is a duplicate of [Squash my last X commits together using Git](https://stackoverflow.com/questions/5189560/squash-my-last-x-commits-together-using-git). — Tim Biegeleisen, Jun 21 '19 at 11:21

score 6 · Accepted Answer · answered Jun 20 '19 at 15:10

6

Allow me to provide the answer that I posted as a comment when the question was marked as a duplicate (and before Tim's answer):

git reset --soft HEAD~10
git commit -m "there, my 10 commits squashed"

Force-pushing (which is fine) as added by Tim is required if you will place this branch on other remote branches if you had already pushed the removed revisions there.

answered Jun 20 '19 at 15:10

eftshift0

26,375
3
36
60

1

Right. And in general, it's also useful to get to know the "git rebase -i HEAD~10" (interactive) command, which lets you squash, reorder, drop, etc., revisions. It can come in very handy. – Nadav Har'El Jun 20 '19 at 16:22
I won't go as far as to say that `rebase -i` is not needed.... I would use it depending on the situation.... in this case, I would just use my recipe. – eftshift0 Jun 20 '19 at 16:25
Can I mention commit id, from where I want to squash.? – Jitesh Mohite Jun 21 '19 at 11:03
On the reset --soft? Sure, you can use the ID. – eftshift0 Jun 21 '19 at 14:06

score 2 · Answer 2 · answered Jun 20 '19 at 14:20

2

Perhaps the easiest approach here would be to just do a soft reset, going back 10 commits, and then recommitting:

git reset --soft HEAD~10
git commit -m 'squash 10 commits into just one commit'
git push --force origin your_branch

The idea behind this soft reset trick is that we move the HEAD pointer back 10 commits. This places all the changes from your 10 commits into the stage, while leaving the working directory the same. Then, when you recommit, you effectively create a commit containing the work from all 10 commits.

Note that a force push would generally be required, since this option rewrites the history of your branch.

Another approach would be to use an interactive rebase, and choose squash for the commits you want to squash. This option would also rewrite history.

answered Jun 20 '19 at 14:20

Tim Biegeleisen

502,043
27
286
360

Can I mention commit id, from where I want to squash. ? – Jitesh Mohite Jun 21 '19 at 11:03
If you want to squash a certain range of commits which ends _before_ the HEAD commit, then a soft reset won't be an option and you'll have to use an interactive rebase. – Tim Biegeleisen Jun 21 '19 at 11:05
That's what I want actually, can you provide the answer for it, I will accept then – Jitesh Mohite Jun 21 '19 at 11:06

jmunsch · Answer 3 · 2019-07-12T14:54:38.497

What this does is creates a new branch "your-branch-3" and then checks out each file from the other branch "your-branch-2":

git checkout master
git checkout -b your-branch-3
git diff your-branch-2 --name-only | xargs -I{} git checkout your-branch-2 -- {}

--name-only

output only the file name paths

xargs -I

Replace occurrences of {} in the initial-arguments with names read from standard input.

git-checkout branch --

double dash meaning https://git-scm.com/docs/git-checkout#_argument_disambiguation

Remove all commits but keep all file changes

3 Answers3