Git Reset

The git reset command is used to undo changes in your Git history. It can move your branch pointer, remove commits, and unstage or discard changes depending on how you use it.

Introduction
- When to Use Reset
Reset Command Options
- Commit References

Introduction

git reset works by changing where your current branch is pointing. You can use it to:

Unstage files that were added with git add
Remove one or more recent commits
Discard changes in your working directory

The reset command is useful for fixing mistakes or cleaning up your Git history. However, it should be used carefully, especially when working with commits that have already been pushed. Improper use of this command can result in data loss.

When to Use Reset

git reset is useful in the following cases:

Unstage files: You ran git add, but now you want to remove the file from staging without deleting your changes.
Modify your commit history: You committed too early or too often and want to rewrite recent commits.
Undo a commit: You made a mistake and want to erase the commit as if it never happened (locally).

Reset Command Options

Here are the most common ways to use git reset:

Unstage a file
```
 $ git reset <file>
```
This removes the file from the staging area, but keeps your changes in the working directory.
Soft reset
```
 $ git reset --soft <commit>
```
Undoes recent commits. Moves the branch pointer back to the given commit but keeps all your changes staged.
Mixed reset
```
 $ git reset --mixed <commit>
```
Undoes commits and unstage changes. Moves the branch pointer back and unstages the changes, but keeps them in your working directory.
Hard reset
```
 $ git reset --hard <commit>
```
Discards commits and changes. This deletes the commit in the history after the given commit and all changes in your working directory.

Hard resets cannot be undone easily. Use this command with extreme caution. Check out this site on git reflog if for more information.

Commit References

In each of the command options above, <commit> represents a commit reference. You can specify a specific commit SHA or branch name here.

Here are some common types of commit references:

HEAD
Refers to your current position in the repository (usually the latest commit on your branch).
HEAD~n
Refers to the commit n steps before the current one. For example, HEAD~1 is one commit ago, HEAD~3 is three commits ago.
HEAD^
Refers to the parent of the current commit, or the commit which a merge was applied to. Similar to HEAD~1, but used when dealing with merge commits. For example, HEAD^2 refers to the second parent of a merge.
Commit SHA
Every Git commit has a unique hash (SHA). This can be found in your repository’s commit history, or in your terminal using:
```
$ git log
```
Included in this output will be a line starting with commit - the text after that is the commit SHA. An example SHA might be d94b5f7ec7c6d7602c78a5e7b8a5b8c54d093eda.
Branch name You can also use a branch name to reference the latest commit on that branch (e.g., main or my-branch).

← Git Rebase ↑ To Advanced SSH Key →