Reset

The reset command is used to change the state of the Git repository or undo commits.

Note: This command is ideal for undoing uncommited changes made in a private, local repository. For undoing changes in a public, remote repository, the revert command is recommended.

Syntax

git reset <mode-option> <commit-reference>

This is run in the terminal. The <mode-options> and <commit-reference> are discussed in more detail below.

Mode Options

The <mode-options> refer to how far reset will go when rolling back changes to a previous commit, including:

  • Where the HEAD is pointing towards (usually done with just git reset).
  • Whether the staging area or index, reflects the commit the HEAD is now pointing towards.
  • Whether the working tree is rolled back to reflect the changes reset in the HEAD and staging area.

More specifically, these modes include:

  • --soft: This rolls back to an earlier commit by moving the HEAD pointer towards it but leaving the staging area and working tree untouched, allowing for new commits to be made.
  • --mixed: In addition to moving the HEAD pointer to an earlier commit, the staging area is cleared to reflect the changes made in that commit (this runs by default).
  • --hard: This goes one step further and resets the working tree to reflect the previous commit reflected in the staging area and the HEAD pointer.

If, for example, an error was made in a text file, example.txt, and the changes were accidentally added and committed, git reset can be used to go back to the state before that commit was made.

Referencing Commits

The commit-reference refers to a commit’s unique hash, or save point, that was 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: 05df67f9066c8ddd95c8d7bb2137acfb8b18e167 -> 05df67f

git reset can be used with either the commit hash or with the HEAD keyword, which refers to the commit being viewed on the currently checked-out branch.

Alternatively, a filename can be used in place of the commit-reference to undo a git add for a file that wasn’t meant to be staged for commit.

Example

This is what the terminal would look like after creating a commit by accident on the main branch and running a git status check:

On branch main
nothing to commit, working tree clean

The text above indicates the following:

  • The main branch is up to date, with the HEAD pointing towards the most recent commit.
  • There is nothing to commit in the staging area.
  • New changes haven’t been made yet in the working tree, hence why it is “clean”.

To set the HEAD back by one commit as well as clear the staging area, one of the following commands can be run:

git reset HEAD~1
git reset --mixed HEAD~1

Since the --mixed mode runs by default, both of the commands are identical in function. This will do the following:

  • It will move the HEAD pointer back by one (~1) commit.
  • The staging area will be cleared of changes.
  • The overall state of the main branch is set to before changes in example.txt were added to the staging area for the previous commit.

If git status is run once more, this should appear on the terminal:

On branch main
Changes not staged for commit:
(use "git add <file>..." to update what will be committed)
(use "git restore <file>..." to discard changes in working directory)
modified: example.txt
no changes added to commit (use "git add" and/or "git commit -a")

Contributors

Interested in helping build Docs? Read the Contribution Guide or share your thoughts in this feedback form.

Learn Git on Codecademy