Revert to original version of a file in git (before even adding)

Question

I am new in Git and my experience is from Clearcase.
In Clearcase when I wanted to modify a file, I checked-out the file to make it writable, modified it and if I was not happy with the changes I simply did an undo checkout going back to the original version of the file.
With git it is like black box.
I can start writing to my file (it is not read only, right?) and then I am supposed to do add followed by commit.
But what if I am not happy with my changes and decided not to even add?
How do I go back to the original file removing my modifications essentially?

`git checkout HEAD -- path/to/file` – jthill May 12 '13 at 21:39 — jthill, May 12 '13 at 21:39

score 4 · Answer 1 · answered May 12 '13 at 19:17

4

The git status output tells you how to do so:

# Changes not staged for commit:
#   (use "git add <file>..." to update what will be committed)
#   (use "git checkout -- <file>..." to discard changes in working directory)

Use git checkout -- <file> to discard changes even before add.

If you have already add the file and then you find it's not satisfactory,

# Changes to be committed:
#   (use "git reset HEAD <file>..." to unstage)

Use git reset HEAD <file> to unstage.

answered May 12 '13 at 19:17

Yang

7,712
9
48
65

`git status` reports on the whole code base or the current directory? Confused on this. I tried to run it in a large code base and seemed to take forever and I killed it – Cratylus May 12 '13 at 19:21
@Cratylus You may add files/folders (e.g. data files, binary files, etc) to a `.gitignore` file to stop tracking them. Typically a code base should not contain too many changed files. – Yang May 12 '13 at 19:30
@Cratylus What are the files `git` is reporting? – Yang May 12 '13 at 19:38
Yang:I stopped `git status` after waiting for ~20secs. I assumed that it scanned all the repository to see the status. May be I am thinking this in the `Clearcase` model I have in my mind and I shouldn't – Cratylus May 12 '13 at 19:42
@Cratylus Can you use `git status | more` to see page by page what files it's reporting? – Yang May 12 '13 at 19:43
Thank you!I will try this as soon as I can. +1 – Cratylus May 12 '13 at 19:47
if `git status` takes >20 secs, `git status|more` will not be any better. you seem to have quite an unoptimized repo there. – eis May 12 '13 at 20:22

score 2 · Answer 2 · edited May 23 '17 at 10:25

May be I am thinking this in the Clearcase model I have in my mind and I shouldn't

Yes: see the main differences between ClearCase and Git here.

Check "Git Status Takes a Long Time to Complete" to see why git status takes a long time: it shouldn't exceed a few seconds at most (unless there is a pager issue).

Remember your git repo is much smaller than a ClearCase Vob. I have Vobs which are huge (several Terabytes). A git repo shouldn't exceed a few Megabytes, in order to be cloned around efficiently.
Make sure your current repo doesn't represent what was a huge ClearCase Vob, with potentially multiple projects in it.

Regarding your initial question, I like the unadd/unstagealias defined in "Undo 'git add' before commit".

git config --global alias.unadd 'reset HEAD --'
git config --global alias.unstage 'reset HEAD --'

Revert to original version of a file in git (before even adding)

2 Answers2