Published Jun 21, 2023
Contribute to Docs

In Git, the revert command is used to create a new commit that will undo any changes made in a previous commit. This can be considered as the safe undo command because it keeps previous history. The command is best used when working with others and there’s a need to preserve the commit history.

Note: This command can be mistaken for the reset command, which is a more dangerous undo command because it alters existing hitory.


git revert <commit-reference>

The commit-reference is the unique hash of a commit that is generated after creation. This hash is a long string that is a mix of characters and numbers that is usually represented by a shorter version: a63b1329066c8ddd95c8d7bb201bacfb8b18e167 -> a63b132

git revert can be used with the commit hash (as seen below) or with the HEAD keyword (as seen below), which refers to the commit being viewed on the currently checked-out branch.


Lets say there are four commits with different files being added for each commit. To view the commits, use git reflog:

$ git reflog
0caf1ae (HEAD -> master) HEAD@{0}: commit: Add file 4
ffd3d9c HEAD@{1}: commit: add file 3
f85ef36 HEAD@{2}: commit: add file 2
fc3980d HEAD@{3}: commit (initial): add file 1

The above shows that each commit has a new file added. File 1 is the first commit and file 4 is the most recent commit.

The writer for file 2 mentions that file 2 is no longer needed. To remove the file, while preserving all the other commits, the revert command is used with the hash for the commit in which file 2 was added:

git revert f85ef36

Another option to remove file 2 is with the HEAD keyword. Looking at the same log as before, it shows the commit for file 2 is located at HEAD@{2}. File 2 can then be removed with the following line:

git revert HEAD~2

By running the git revert command, this will then open a prompt to edit the message before committing the revert. Once the message has been saved, check the logs to view the status of file 2:

$ git reflog
f85ef36 (HEAD -> master) HEAD@{0}: revert: Revert "add file 2"
0caf1ae HEAD@{1}: commit: Add file 4
ffd3d9c HEAD@{2}: commit: add file 3
f85ef36 HEAD@{3}: commit: add file 2
fc3980d HEAD@{4}: commit (initial): add file 1

The log still shows the original commit of “add file 2”, but it has moved forward and deleted file 2 with the most recent commit. The end result is file 1, file 3, and file 4 are preserved, while file 2 no longer exists. By using git revert, a new commit is created to delete the specific commit while preserving the previous history.

The revert command also has options that can be added such as:

  • -e or --edit: This is the default option where it opens a prompt to edit the commit message before committing the revert.
  • --no-edit: This bypasses the prompt for a commit message edit.
  • -n or --no-commit: This will revert changes to the previous state without completing the commit.

All contributors

Looking to contribute?

Learn Git on Codecademy