Git Staging: How to Stage, Unstage and Commit Changes

Git came with exciting features like git staging and git unstaging files or Committing changes to a branch. As you know, We developers add, modify or delete files daily. But we have to commit to a staging environment whenever we complete any task or work.

Let’s understand staging, unstaging, committing work, add changes in Hunk step by step.

How to Stage All Changes to Files

git add -A

git add . # Version >= 2.0

In version 2.x, git add . will stage all changes to files in the current directory and all its subdirectories. However, in

1.x It will only stage new and modified files, not deleted files.

Use git add -A, or its equivalent command git add --all, to stage all changes to files in any version of git.

Unstage a file that contains changes

git reset <filePath>

How to Add changes by hunk

You can see what “hunks” of work would be staged for commit using the patch flag:

git add -p

or

git add --patch

This opens an interactive prompt that allows you to look at the diffs and let you decide whether you want to include them or not.

Stage this hunk [y,n,q,a,d,/,s,e,?]?

This makes it easy to catch changes that you do not want to commit.

You can also open this via git add --interactive and selecting p.

Interactive add in Git

git add -i (or --interactive) will give you an interactive interface where you can edit the index, to prepare what you want to have in the next commit. You can add and remove changes to whole files, add untracked files and remove files from being tracked, but also select subsections of changes to put in the index, by selecting chunks of changes to be added, splitting those chunks, or even editing the diff. Many graphical commit tools for Git (like e.g. git gui) include such a feature; this might be easier to use than the command line version.

It is very useful (1) if you have entangled changes in the working directory that you want to put in separate commits and not all in one single commit and (2) if you are in the middle of an interactive rebase and want to split too large a commit.

$ git add -i
              staged        unstaged path 
1:         unchanged        +4/-4 index.js
2:             +1/-0       nothing package.json

*** Commands ***
  1: status  2: update  3: revert  4: add untracked
  5: patch  6: diff  7: quit  8: help
What now>

The top half of this output shows the current state of the index broken up into staged and unstaged columns:

• index.js has had 4 lines added and 4 lines removed. It is currently not staged, as the current status reports “unchanged.” When this file becomes staged, the +4/-4 bit will be transferred to the staged column and the unstaged column will read “nothing.”
• package.json has had one line added and has been staged. There are no further changes since it has been staged as indicated by the “nothing” line under the unstaged column.
• The bottom half shows what you can do. Either enter a number (1-8) or a letter (s, u, r, a, p, d, q, h).
• status shows output identical to the top part of the output above.
• update allows you to make further changes to the staged commits with additional syntax.
• revert will revert the staged commit information back to HEAD.
• add untracked allows you to add file paths previously untracked by version control.
• patch allows for one path to be selected out of an output similar to the status for further analysis.
• diff displays what will be committed.
• quit exits the command.
• help presents further help on using this command.

How to show Staged Changes

To display the hunks that are staged for commit:

git diff --cached

Staging A Single File in Git

To stage a file for committing, run

git add <filename>

Stage deleted files in Git

git rm filename

To delete the file from git without removing it from the disk, use the --cached flag

git rm --cached filename