What is the difference between 'git pull' and 'git fetch'?

Question

What are the differences between git pull and git fetch?

I found this well written article about git fetch and git pull it's worth the reading: http://longair.net/blog/2009/04/16/git-fetch-and-merge/ — Marcos Oliveira, Sep 16 '10 at 06:57
Our alternative approach has become `git fetch; git reset --hard origin/master` as part of our workflow. It blows away local changes, keeps you up to date with master BUT makes sure you don't just pull in new changes on top on current changes and make a mess. We've used it for a while and it basically feels a lot safer in practice. Just be sure to add/commit/stash any work-in-progress first ! — Michael Durrant, May 04 '14 at 14:32
Make sure you know how to use git stash correctly. If you're asking about 'pull' and 'fetch' then maybe 'stash' will also need explaining... — Henry Heleine, Dec 09 '14 at 20:09
Lots of folks coming from Mercurial keep using "git pull", thinking it's an equivalent for "hg pull". Which it's not. Git's equivalent of "hg pull" is "git fetch". — Serge Shultz, Jun 29 '15 at 10:15
git fetch command fetching updated code with branch and also will get newly added branches in your local, git pull command fetch only updated code of current branch only — Kartik Patel, Jul 27 '17 at 11:43
@MichaelDurrant FWIW, that is equivalent to `git pull --ff-only`, although I don't remember if it was available in 2014. — jpaugh, May 21 '18 at 21:16
To note, `git fetch` has no correspondent in traditional source control systems (that I'm aware of). The main goal of using `git fetch` (to "see what happened in remote") is achieved e.g. in SVN with a `merge --dry-run`. It should be part of the basic workflow, that before you pull, you get an idea of what changes you are pulling, particularly **those changes that would conflict with your local ones**. — haelix, Sep 06 '18 at 18:16
A very good way of understanding them is understanding the various `refs` git stores. That helps you to understand how the isolation mechanism between them works and why `fetch` is safe vs. `pull` which is aggressive. For more see [here](https://stackoverflow.com/questions/292357/what-is-the-difference-between-git-pull-and-git-fetch/54657340#54657340) — Honey, Feb 12 '19 at 19:30
https://www.freecodecamp.org/news/git-fetch-vs-pull/#:~:text=git%20fetch%20is%20the%20command,changes%20from%20the%20remote%20repository. — Shiwangini, Jul 11 '20 at 19:24

score 10457 · Accepted Answer · edited May 14 '21 at 11:26

10457

In the simplest terms, git pull does a git fetch followed by a git merge.

You can do a git fetch at any time to update your remote-tracking branches under refs/remotes/<remote>/. This operation never changes any of your own local branches under refs/heads, and is safe to do without changing your working copy. I have even heard of people running git fetch periodically in a cron job in the background (although I wouldn't recommend doing this).

A git pull is what you would do to bring a local branch up-to-date with its remote version, while also updating your other remote-tracking branches.

From the Git documentation for git pull:

In its default mode, git pull is shorthand for git fetch followed by git merge FETCH_HEAD.

edited May 14 '21 at 11:26

Mateen Ulhaq

18,406
13
75
112

answered Nov 15 '08 at 09:52

Greg Hewgill

828,234
170
1,097
1,237

347

"A "git pull" is what you would do to bring your repository up to date" – Albert Nov 10 '09 at 12:13
204

@Albert: Yeah, it's weirdly worded. `git pull` will always merge into the **current branch**. So you select which branch you want to pull *from*, and it pulls it into the current branch. The *from* branch can be local or remote; it can even be a remote branch that's not a registered `git remote` (meaning you pass a URL on the `git pull` command line). – intuited Jun 06 '10 at 10:10
13

Is a "git push" also a "git fetch" (in the other direction) followed by a "git merge"? – Ellen Spertus Mar 16 '11 at 23:51
133

@espertus: No. Pushing never automatically does a merge. The user is expected to pull, resolving any merge conflicts locally, *then* push back to the remote. – Greg Hewgill Mar 17 '11 at 00:41
35

If I am at `/home/alice/` and do `git fetch /home/bob`, what parameters should I pass to the subsequent `git merge` ? – ripper234 May 27 '11 at 19:38
9

Why specifically don't you recommend something like a cron job that does a "git fetch" automatically for you? No bad side effects come to mind, and it'd be nice to have my remote branches updated without having to think about it. – Bryan Henry Oct 27 '11 at 02:05
11

@Sbrocket: I don't like surprises, especially not while working with my source control system. Changing branches from under me (even just remote branches) while working could be disorienting. – Greg Hewgill Oct 27 '11 at 18:29
112

Note to people learning Git: `pull` can't actually be emulated by a `fetch` plus a `merge`. I just fetched a change where only a remote branch pointer changes, and `merge` refuses to do anything. `pull`, on the other hand, fast-forwards my tracking branch. – Roman Starkov Sep 28 '12 at 16:23
13

Lets imagine that I have only a `master` branch. Which EXACT commands (with all the necessary options) should replace `git pull` command? – Paul Jan 25 '13 at 15:16
@espertus - FYI if you uses Git Extensions you can do auto-pull on before push if the push is rejected. – Laksitha Ranasingha Jan 31 '13 at 12:36
5

Alternatively, you can update your git repo with remote repo by doing `git fetch` followed by `git rebase origin/master` – abumusamq Feb 07 '13 at 12:16
8

@romkyns I don't quite follow what you mean there. Could you please explain what "I just fetched a change where only a remote branch pointer changes, and merge refuses to do anything." means? – viki.omega9 May 26 '13 at 17:15
2

@dev-inside Or just do `git pull --rebase` – Michael Mior May 26 '13 at 21:06
34

git pull is shorthand for git fetch followed by git merge FETCH_HEAD. – Nikita Apr 15 '14 at 18:51
6

**This operation never changes any of your own local branches under refs/heads** - what does it do then? I still don't understand well. – アレックス Jun 13 '14 at 01:39
5

@AlexanderSupertramp: `git fetch` updates the *remote tracking* branches. These are the branches listed starting with `remotes/` in `git branch -a`, and are a copy of the branches on the remote repository. – Greg Hewgill Jun 13 '14 at 01:41
@GregHewgill, how does it updates them? It seems like you're implying it pushes something to the remote branches. – アレックス Jun 13 '14 at 04:38
6

@AlexanderSupertramp: I'm not saying that at all. Your normal branches are under `refs/heads`. Remote tracking branches are under `refs/remotes`. The `git fetch` command updates your remote tracking branches under `refs/remotes` and does not change anything under `refs/heads`. – Greg Hewgill Jun 13 '14 at 07:16
1

Locally I keep only local branches and edit only them, correct? What do you mean by "updates remotes" branches? – アレックス Jun 13 '14 at 08:00
5

@AlexanderSupertramp: Please ask a new question if you are having trouble understanding Git terminology. The comments section is not the right place to have this discussion. Thanks. – Greg Hewgill Jun 13 '14 at 08:02
3

The key distinction here is refs/remote and refs/heads. heads refers to your local working branch, whereas remote refers to the branch pointing to the remote, with information since your last git fetch origin. – JohnMerlino Jun 28 '14 at 16:22
@romkyns "`pull` can't actually be emulated by a `fetch` plus a `merge`" Actually, yes it can. I do it all the time. – Ajedi32 Aug 20 '14 at 18:23
1

@Ajedi32 some of what `pull` does can be emulated by those two, yes. But not the entire command in all possible circumstances. – Roman Starkov Aug 22 '14 at 14:22
2

@romkyns True. Perhaps I didn't understand the circumstance mentioned in your comment, but fast-forwards can definitely be done with `git merge`. – Ajedi32 Aug 23 '14 at 03:09
"while also updating your other remote-tracking branches" What does this mean? – user2906759 Sep 11 '14 at 12:50
5

@user2906759: Git keeps track, in your local repository, of the state of all the branches on a remote. Google "remote tracking branches" for lots of info about this. – Greg Hewgill Sep 11 '14 at 18:52
Something I've been playing around with. I often 'git fetch' to pull the changes. If you then 'git pull' though it appears that it goes to the network again. Instead, AFTER the fetch, I've been running 'git merge @{u}' which I think merges in the head of the configured tracking branch without going to the network again. Seems faster. Warning - I'm still experimenting! – Damien Sawyer Dec 16 '14 at 20:53
2

Is this the true workings of git pull? Or simply a pseudo-similar behavior? Has anyone looked at the implementation of git pull as really using git fetch and git merge under the hood? Or does git pull has it's own internal logic, which in most situations performs similar operations as to what fetch followed by merge does, but possibly not in all of them? – Didier A. Jan 05 '15 at 16:44
1

You mentioned a cron job to automatically do `git fetch`. I think SourceTree has an option to periodically run `git fetch`, since I seem to notice remote updates in the visual interface even though I didn't run any explicit commands. – Doctor Blue Jan 22 '15 at 14:16
5

Now I understand why `pull` from `master` into the current branch won't update the branch even if `master` has new local commits merged in. It's because the `merge` in `pull` is merging from `FETCH_HEAD` and completely ignoring my local `master`! I wish this important difference was better emphasized in all tutorials which mention `pull` and claim that it is almost the same as `fetch` + `merge`. "Almost" is really important if you are `pull`ing and hoping to do all-in-one - fetch + merge from remote AND merge from local. Surprise - pull does not merge from local branch! – JustAMartin Oct 21 '15 at 19:53
1

Does "update remote-tracking branches" mean that all the branches up on the repository now are available on your local repository? – Jwan622 Mar 25 '16 at 16:16
1

@Jwan622: Yes, that's correct. `git branch -a` will list them all. – Greg Hewgill Mar 26 '16 at 01:06
I saw the same answer in Quora :) Here : https://www.quora.com/Whats-the-difference-between-git-pulland-git-fetch. – Pragyaditya Das Apr 02 '16 at 05:27
1

If you work with a bad internet connection, it may be a good idea to run a cron job to `git fetch`. – noɥʇʎԀʎzɐɹƆ Oct 10 '16 at 22:35
Somewhat vague explanations, even in git's ref docs. Docs say "git pull runs git fetch with the given parameters and calls git merge to merge the retrieved branch heads into the current branch." Sounds like a nightmare if ALL fetched are merged into current branch. – Josef.B Dec 15 '16 at 02:54
1

Git fetch means you are get new repository from git which is created by someone. Git pull means you are getting new code from the git or some of the repository which is pushed by some one. – Chirag Prajapati Nov 20 '17 at 18:27
git pull use for getting updated merged code from repository. git fetch is for getting branches in the repository to exchange branch into a new branch. – Hasib Kamal Mar 19 '18 at 05:10
2

@GregHewgill If fetch only downloads changes from the remote branch and updates the repository data, but leaves the local branch unchanged, what is the point of fetching if the working directory will not show/reflect the changes? My question originally was that how could I see the changes someone else has done and then decide whether I would like to merge them into my working directory (i.e. experiment with other people's changes to make sure it does not break my work) but I am still confused how to do that. Should I just pul and experiment/explore, and if it was problematic, do a hard reset? – Jun 09 '18 at 17:02
1

@Joshua: After fetching, the remote branches are available with names like `origin/master` (instead of just `master`, which is yours). You can view those branches locally to see what's changed before actually merging. – Greg Hewgill Jun 09 '18 at 23:18
The current (see datetime stamp) Visual Studio Code C/C++ extension sends a popup notification asking to perform periodic *git fetch* commands. – bscout11 Feb 05 '20 at 05:33

score 2281 · Answer 2 · edited May 14 '21 at 11:35

2281

When you use pull, Git tries to automatically merge. It is context sensitive, so Git will merge any pulled commits into the branch you are currently working on. pull automatically merges the commits without letting you review them first. If you don’t carefully manage your branches, you may run into frequent conflicts.
When you fetch, Git gathers any commits from the target branch that do not exist in your current branch and stores them in your local repository. However, it does not merge them with your current branch. This is particularly useful if you need to keep your repository up to date, but are working on something that might break if you update your files. To integrate the commits into your current branch, you must use merge afterwards.

edited May 14 '21 at 11:35

Mateen Ulhaq

18,406
13
75
112

answered Aug 18 '11 at 08:53

Mouna Cheikhna

35,154
10
46
68

36

Agreed, great comment. Which is why I hate git pull. When would it ever make sense to let a revision tool make code edits for you? And isn't that what merging two files is doing? What if those two edits are physically separated in the file, but LOGICALLY at odds? – Lee Dixon May 13 '13 at 18:44
I'm not sure if I understand this correctly. Let me know if I'm right: Lets say I have two branches, master and test. test is a branch that I'm working on to experiment something. If I do git fetch, it updates master with the target branch. If I do git pull, it tries to update test with the target branch. Is this right? If not, I think I don't understand what 'local repository' means - I assumed it means my local master. – elexhobby Jun 05 '13 at 19:15
136

@elexhobby short put, `git fetch` only updates your `.git/` directory (AKA: local repository) and nothing outside `.git/` (AKA: working tree). It does not change your local branches, and it does not touch `master` either. It touches `remotes/origin/master` though (see `git branch -avv`). If you have more remotes, try `git remote update`. This is a `git fetch` for all remotes in one command. – Tino Jul 17 '13 at 06:48
26

@Tino yours is really the most important point. People may not know that "remote" branches are actually stored as a bunch of hashes in `.git/refs/remotes/origin/`. – Chris Sep 12 '13 at 21:49
4

Once I understood that everything was going 'though' tracking branches I finally really understood git. Without that knowledge it's just 'magic'. – Michael Durrant Sep 19 '13 at 19:55
15

**When you fetch, Git gathers any commits from the target branch that do not exist in your current branch and stores them in your local repository** - how do I see what was brought from the remote and how do I merge it into my local branches? – アレックス Jun 13 '14 at 01:40
15

@Tino What I still don't understand is... what's the point? Why use fetch if it only updates `.git`? What is the intended benefit and what am I supposed to do after that? – BadHorsie Mar 23 '16 at 12:19
@BadHorsie `git fetch` downloads the changesets. – Tino Mar 30 '16 at 13:40
4

`git pull --ff-only` is usually what I want. Does the mindless stuff automatically, asks before making history complicated. – Nick T Dec 18 '17 at 22:05
3

I find this answer great. For me a good combination to achieve the goal from git fetch is typing 'git status' and you will see something like : "nothing to commit, working tree clean" then you can type 'git fetch' and then type 'git status' and you will see that your branch is "ahead of the remote repository" or smth like that. If you type "git pull" you will merge the branch but with 'git fetch' you only make conscious your local repository if you are ahead or behind with some commits. – eduardo92 Jun 12 '18 at 13:39
1

`git pull` means "fetch remote branches and merge my current branch with whatever random commits have been done in remote branches". That pretty much never makes any sense. `git fetch` followed by `gitk` or `git log --oneline --graph --decorate --date-order` and the maybe `git rebase` or `git merge` makes much more sense. – Mikko Rantalainen Oct 25 '18 at 10:50
1

It's been a long time, but to answer @BadHorsie's question another way, it's useful to fetch, even though it won't affect your working branch, because you might want to perform comparisons against the remote branch, or you might be going off-line - say on an airplane - and know that you will need that remote data when you can't get it. I'm sure there are even more reasons, but these are the two I run into frequently. – Thagomizer Sep 23 '19 at 22:11

score 1270 · Answer 3 · edited Aug 24 '18 at 20:01

1270

It is important to contrast the design philosophy of git with the philosophy of a more traditional source control tool like SVN.

Subversion was designed and built with a client/server model. There is a single repository that is the server, and several clients can fetch code from the server, work on it, then commit it back to the server. The assumption is that the client can always contact the server when it needs to perform an operation.

Git was designed to support a more distributed model with no need for a central repository (though you can certainly use one if you like). Also git was designed so that the client and the "server" don't need to be online at the same time. Git was designed so that people on an unreliable link could exchange code via email, even. It is possible to work completely disconnected and burn a CD to exchange code via git.

In order to support this model git maintains a local repository with your code and also an additional local repository that mirrors the state of the remote repository. By keeping a copy of the remote repository locally, git can figure out the changes needed even when the remote repository is not reachable. Later when you need to send the changes to someone else, git can transfer them as a set of changes from a point in time known to the remote repository.

git fetch is the command that says "bring my local copy of the remote repository up to date."
git pull says "bring the changes in the remote repository to where I keep my own code."

Normally git pull does this by doing a git fetch to bring the local copy of the remote repository up to date, and then merging the changes into your own code repository and possibly your working copy.

The take away is to keep in mind that there are often at least three copies of a project on your workstation. One copy is your own repository with your own commit history. The second copy is your working copy where you are editing and building. The third copy is your local "cached" copy of a remote repository.

edited Aug 24 '18 at 20:01

Ian Ringrose

49,271
50
203
302

answered Mar 31 '13 at 18:43

MikeD

13,208
1
17
9

80

Technically, the local and remote repositories are really one and the same. In Git, a repository is a [DAG](http://en.wikipedia.org/wiki/Directed_acyclic_graph) of commits pointing to their parents. Branches are, technically, nothing more than meaningful names of commits. The only difference between local and remote branches is that remote ones are prefixed with `remoteName/` [Git from the ground up](http://newartisans.com/2008/04/git-from-the-bottom-up/) is a very good read. Once you get an understanding of how Git works - and it's beautifully _simple_, really - everything just makes sense. – Emil Lundberg Aug 14 '13 at 09:51
13

Thanks much for the explanation. I didn't really understand until now that Git was designed so you didn't have to have a central repository. Everyone always says "DVCS" when describing Git, but as a relatively new programmer, that means nothing to me. I've never **seen** a CVCS, and I've also never not worked with a cental remote repository when collaborating with others (i.e. Github), so until now I've yet to understand what made Git special. – Brian Peterson Aug 15 '13 at 02:17
8

So, based on this, why ISN'T it a good idea to git-fetch with a cron job? Always keeping a copy of the remote you're working with on your local machine seems like a good idea. In fact, I feel like writing a script that checks to see if I've updated my remote in the past 24 hours and linking it up with a udev hook for internet connection. – Brian Peterson Aug 15 '13 at 02:23
Silly question - Are these "three copies" maintained by Git and I never see it? or I have downloaded one copy and created two branches? – Nabheet Sep 14 '13 at 12:47
2

@Nabheet: I believe the additional copies are not stored as ordinary files on your file system, but instead as alternate histories (changesets) inside your git repository files. – Keen Sep 17 '13 at 02:55
26

One reason why it ISN'T a good idea to have a cron job: often when working on either a new ticket, or on updates to a branch, I like to see the changes being fetched. If the changes don't come in during a fetch, I'll be more confident in asking my fellow programmer 'hey did you push?'. I also get a sense of how much 'churn' in the repository since I last fetched. This also help to give me sense of the amount and speed of changes currently being made to this repository. – Michael Durrant Jul 23 '14 at 11:47
5

@Nabheet Thing is that, Git is content-oriented. It stores data only once, and points to it multiple times. That's why in Git, even multiple commits on top of an original don't affect the size of the repo much, since most of the objects are the same. – cst1992 Nov 05 '16 at 09:43
this is the most clarifying answer. I like to keep it simple but I would just add some technical names in parenthesis like Head, index and working tree/dir along with the simple explanations/analogies (e.g. where I keep my own code) – Luis Martins Feb 24 '18 at 02:52
I have a small typo there. So `git pull` first brings my local copy of the remote repository up to date just like `git fetch`, then it merges those changes into my local repository where I have my own code. Imagine I have made some local commits, and someone else has made commits and pushed it to a remote repository. At first, I know nothing about the commits the other person made. When I execute `git fetch`, I get a copy of the other person's work stored locally. Now I can compare it to mine. When I execute `git pull` I can merge their changes with my changes in my local workspace. – MikeD Mar 14 '18 at 18:52
@MikeD >> _Git was designed to support a more distributed model with no need for a central repository_. Read - to support people still _committing often_ during long flights (during which there is no Internet). – haelix Sep 06 '18 at 18:23
This answer should be combined with the next down by Contango containing the image of the relationship. Then appended to the top answer, by Greg Hewgill, that gives the briefest and clearest explanation of the consequences of pull vs fetch. The order of the answers so far though is, fortunately, effectively that. – GG2 Oct 01 '18 at 18:06

score 953 · Answer 4 · edited Mar 19 '18 at 19:21

953

Here is Oliver Steele's image of how all it all fits together:

enter image description here

If there is sufficient interest, I suppose I could update the image to add git clone and git merge...

edited Mar 19 '18 at 19:21

nmichaels

45,418
12
95
127

answered Jun 09 '15 at 13:30

Contango

65,385
53
229
279

166

An updated image with `git clone` and `git merge` would be very helpful! – MEMark Sep 08 '15 at 14:23
20

Yes, please add `git merge` - it should clearly show that `merge` called separately is NOT the same as calling `pull` because `pull` is merging from remote only and ignores your local commits in your local branch which is tracking the remote branch being pulled from. – JustAMartin Oct 21 '15 at 19:57
12

A picture is worth a thousand words! Is the updated image with clone and merge data flow ready somewhere? Any other data flow besides what's already in the diagram? – shikhanshu Nov 25 '15 at 00:27
10

@Contango please add clone and merge. Would be helpful for newbies like me. – rents Jan 15 '16 at 16:02
11

There are two diagrams showing clone and merge in other answers (below) by th3sly and thedarkpassenger. – intotecho Aug 12 '16 at 00:42
5

Does rebase come from the remote repository? I think that rebase is an operation over the local repository: https://git-scm.com/docs/git-rebase – Edward Ross Sep 26 '16 at 09:55
3

Rebase works on your local repository. The confusion is that the most common use case is to rebase your work on top of something that "came from" a remote repository. – MikeD Jun 20 '17 at 20:02
5

This 'pull or rebase' is wrong. Because when you pull, you actually perform a fetch and than a merge. When you fetch, you get new changes from the remote branch to your local branch, but, and that's the important part, it's not added to your files yet. It's like the changes are parked away at some place. The changes will be added to your files, if you merge them. Than you changes are 'melted' to your files and committed. – AndaluZ Sep 13 '17 at 09:36
1

Agree with @EdwardRoss, @MikeD and @AndaluZ: Rebase is a local repo to working copy operation. IMEHO, I'd change this so that Pull is remote repo to both local repo and working copy; Fetch is remote repo to local repo, and both Merge and Rebase are local repo to working copy. Part of the confusion ensues because git rebase is like git checkout, it's one command that has several different effects. e.g. `git checkout` can have the effect of either `svn switch` or `svn revert` (and others) depending on how it's used. – dgnuff Mar 27 '18 at 20:10
7

Still waiting on the updated image 4 years later :) – JOATMON Oct 10 '19 at 18:50

score 514 · Answer 5 · edited Jul 16 '19 at 19:49

514

One use case of git fetch is that the following will tell you any changes in the remote branch since your last pull... so you can check before doing an actual pull, which could change files in your current branch and working copy.

git fetch
git diff ...origin

See: https://git-scm.com/docs/git-diff regarding double- and triple-dot syntax in the diff command

edited Jul 16 '19 at 19:49

rkedge

338
2
11

answered May 07 '10 at 19:23

mepster

5,173
1
13
2

10

why not `git diff ..origin`? – Erik Kaplun Feb 12 '12 at 23:47
5

git diff origin and git diff ..origin seem to work but not this weird ... stuff – Marc Jan 08 '13 at 19:32
19

@Compustretch There was not supposed to be a space. `git diff ...origin` is equivalent to `git diff $(git-merge-base HEAD origin) origin` (see the `git diff [--options] ... [--] […]` section of https://www.kernel.org/pub/software/scm/git/docs/git-diff.html#_description), which is different from `git diff origin`; `git diff ...origin` is conceptually the changes made in `origin` since the current branch branched from `origin`, while `git diff origin` includes also the reverse of the changes made in the current branch since it branched from `origin`. – Max Nanasy Aug 01 '13 at 19:34
3

none of the .. commands worked for me (on Windows), but `git diff origin/master` works, as mentioned below – Brian Burns Feb 20 '14 at 09:07
same here using git 2.0.0 on OSX. None of these commands worked. Have they been deprecated? – K.-Michael Aye Jun 26 '14 at 02:39

score 388 · Answer 6 · edited Jan 02 '16 at 04:52

388

It cost me a little bit to understand what was the difference, but this is a simple explanation. master in your localhost is a branch.

When you clone a repository you fetch the entire repository to you local host. This means that at that time you have an origin/master pointer to HEAD and master pointing to the same HEAD.

when you start working and do commits you advance the master pointer to HEAD + your commits. But the origin/master pointer is still pointing to what it was when you cloned.

So the difference will be:

If you do a git fetch it will just fetch all the changes in the remote repository (GitHub) and move the origin/master pointer to HEAD. Meanwhile your local branch master will keep pointing to where it has.
If you do a git pull, it will do basically fetch (as explained previously) and merge any new changes to your master branch and move the pointer to HEAD.

edited Jan 02 '16 at 04:52

stites

3,913
5
28
43

answered May 11 '12 at 18:37

Gerardo

6,461
3
19
15

15

origin/master is a local branch that is a COPY of master on origin. When you fetch, you update local:/origin/master. Once you really grok that everything in git is a branch, this makes a lot of sense and is a very powerful way to maintain different changesets, make quick local branches, merge and rebase, and generally get a lot of value out of the cheap branching model. – cam8001 May 28 '13 at 16:00
3

Still confusing. I thought `git fetch` was to literally download changes on the remote repo into your local repo, but NOT commit them - ie, they still need to be added/committed to your local repo. – krb686 Feb 26 '15 at 14:57
3

fetch only pulls from remote/origin (github) to your local origin. But it doesn't merge it to your actual working files. if you do a pull it will fetch and the merge to your current working files – Gerardo Feb 26 '15 at 19:45
what if you're not happy with the changes after you git fetch? what to do next? – Ahmad Dec 28 '20 at 23:42
@Ahmad git fetch fetches data from the remote repository. If you are not happy with the some commits that other collaborators pushed, then do a `fetch` it. `Merge it` . Undo the commits and the `push` back to remote repository – Jdeep Apr 05 '21 at 10:14

score 246 · Answer 7 · edited Jan 26 '21 at 21:26

Briefly

git fetch is similar to pull but doesn't merge. i.e. it fetches remote updates (refs and objects) but your local stays the same (i.e. origin/master gets updated but master stays the same) .

git pull pulls down from a remote and instantly merges.

More

git clone clones a repo.

git rebase saves stuff from your current branch that isn't in the upstream branch to a temporary area. Your branch is now the same as before you started your changes. So, git pull -rebase will pull down the remote changes, rewind your local branch, replay your changes over the top of your current branch one by one until you're up-to-date.

Also, git branch -a will show you exactly what’s going on with all your branches - local and remote.

This blog post was useful:

The difference between git pull, git fetch and git clone (and git rebase) - Mike Pearce

and covers git pull, git fetch, git clone and git rebase.

UPDATE

I thought I'd update this to show how you'd actually use this in practice.

Update your local repo from the remote (but don't merge):
```
 git fetch 
```
After downloading the updates, let's see the differences:
```
 git diff master origin/master 
```
If you're happy with those updates, then merge:
```
 git pull
```

Notes:

On step 2: For more on diffs between local and remotes, see: How to compare a local git branch with its remote branch?

On step 3: It's probably more accurate (e.g. on a fast changing repo) to do a git rebase origin here. See @Justin Ohms comment in another answer.

See also: http://longair.net/blog/2009/04/16/git-fetch-and-merge/

Sounds to me like if someone just wants the local code to reflect "the tip", they should use `git clone`. I put the tip in quotes, as I assume it would mean whatever master is and what someone would "Download as zip" from github.com — Chris K, Sep 12 '13 at 08:27
what if you're not happy with the changes after you git fetch? what to do next? — Kugutsumen, Mar 24 '15 at 06:06
Your paragraph on rebase was just what I was looking for. The whole idea about zeroing out of everything, updating from remote, then **replaying your changes on top of previous commits** that happened while you were working. Perfect explanation assuming it's correct. ;) — coblr, Mar 03 '16 at 00:01

score 239 · Answer 8 · answered Jan 25 '16 at 17:28

239

Sometimes a visual representation helps.

answered Jan 25 '16 at 17:28

thedarkpassenger

6,049
2
32
52

20

I think the picture got to show that it affects the local repo too. That's is, Git pull is a combination of affecting the local repo and working copy. Right now it seems it just affect the working copy. – nonopolarity Feb 14 '16 at 18:51
11

@太極者無極而生 Agreed -- this image is pretty misleading, because it makes it look like `git pull` is _skipping_ the fetch, which of course is inaccurate. – forresthopkinsa Aug 20 '16 at 15:33
9

whats a difference between a 'Local Repository' and a 'Working Copy'? Aren't they both local on the computer? – theITvideos Nov 16 '17 at 03:12
1

What is use of git fetch then? how to see what difference is there in local repository and working copy? – Vikash Mar 13 '19 at 06:56
3

@theITvideos No, it is not. A local repository is where your code goes(from working repository) when you commit . (It goes to remote repo when you push). – Vikash Mar 13 '19 at 06:59
this post has lots of upvotes but the picture is misleading. Pull is skipping local repo which is not true. https://codeahoy.com/img/git-pull-vs-fetch.png – goblinjuice Feb 02 '20 at 14:35

score 188 · Answer 9 · answered Nov 15 '08 at 09:52

188

git-pull - Fetch from and merge with another repository or a local branch
SYNOPSIS

git pull   …
DESCRIPTION

Runs git-fetch with the given parameters, and calls git-merge to merge the 
retrieved head(s) into the current branch. With --rebase, calls git-rebase 
instead of git-merge.

Note that you can use . (current directory) as the <repository> to pull 
from the local repository — this is useful when merging local branches 
into the current branch.

Also note that options meant for git-pull itself and underlying git-merge 
must be given before the options meant for git-fetch.

You would pull if you want the histories merged, you'd fetch if you just 'want the codez' as some person has been tagging some articles around here.

answered Nov 15 '08 at 09:52

Vinko Vrsalovic

244,143
49
315
361

5

Very interesting, but I can't really see a use case where you want "just the code". Et what happen with your code when you fetch? Is it erased? What happen whith the remote changes? How does it goes into your repo whithout erasing your code if you don't merge? – e-satis Mar 27 '10 at 16:21
11

@e-satis: The remote branch is also stored locally on your machine. So when you do `git fetch` it fetches changes from the repository and updates your local remote branch. It does not affect your local branch which tracks the local remote branch, so does not affect your working copy. Now, when you do a `merge` it will merge the fetched changes with your local branch. – jeffreyveon Oct 31 '11 at 04:23
A simple use case for the fetch command: perform time consuming operations involving other people's recent commits, such as a merge or a code review, accessing only your up-to-date local repository without network connectivity requirements, because you previously used fetch to download everything you need quickly (e.g. while you are visiting some other developer and connected to some other repository's network). The pull command would download the same commits, but the merging it performs can be undesirable. – Lorenzo Gatti Sep 19 '13 at 10:25

score 170 · Answer 10 · edited Mar 15 '15 at 22:20

170

You can fetch from a remote repository, see the differences and then pull or merge.

This is an example for a remote repository called origin and a branch called master tracking the remote branch origin/master:

git checkout master                                                  
git fetch                                        
git diff origin/master
git rebase origin master

edited Mar 15 '15 at 22:20

Olli

1,208
15
31

answered Mar 21 '11 at 11:07

Antonio Bardazzi

2,620
1
16
19

35

You probably want to skip the pull and just do a "git rebase origin" as the last step since you already fetched the changes. The reason is that someone could have pushed changes in the time since you did the fetch and these would not have been in fetch that you did the diff review on. – Justin Ohms Aug 31 '12 at 20:02

jfmercer · Answer 11 · 2018-09-24T15:28:49.747

166

The short and easy answer is that git pull is simply git fetch followed by git merge.

It is very important to note that git pull will automatically merge whether you like it or not. This could, of course, result in merge conflicts. Let's say your remote is origin and your branch is master. If you git diff origin/master before pulling, you should have some idea of potential merge conflicts and could prepare your local branch accordingly.

In addition to pulling and pushing, some workflows involve git rebase, such as this one, which I paraphrase from the linked article:

git pull origin master
git checkout foo-branch
git rebase master
git push origin foo-branch

If you find yourself in such a situation, you may be tempted to git pull --rebase. Unless you really, really know what you are doing, I would advise against that. This warning is from the man page for git-pull, version 2.3.5:

This is a potentially dangerous mode of operation. It rewrites history, which does not bode well when you published that history already. Do not use this option unless you have read git-rebase(1) carefully.

edited Sep 24 '18 at 15:28

answered May 15 '11 at 20:53

jfmercer

3,313
3
26
35

2

@JustinOhms If `git pull --rebase` is not the right thing in the given situation, is it right if it is done in two steps? If it is the right thing to do, what is extra the benefit to doing it in two steps? – Kaz May 23 '13 at 21:56
@Kaz - because the rebase is not automatic. Fetching the changes first allows you to make the judgement call. It doesn't fix the problem with rebasing history you've already pushed. It will allow you to see if it is safe to rebase changes you have not already pushed. – Justin Ohms May 24 '13 at 21:11
2

@JustinOhms How would you decide whether it is safe to rebase changes? I would just try git rebase, and backtrack if it made a mess, in which case I might as well do git pull --rebase. But maybe you have some other way? – Kaz May 25 '13 at 06:14
3

@KaZ gitk allows you to see the branch structure visually. It will show your the position of your local head, remotes, and your branch structures in relation to what you have fetched. This way you can ensure that you are not rebasing fetched changes that are based on an ancestor that is prior to what you have already pushed to your remote(s). – Justin Ohms May 28 '13 at 19:18
Use `rebase` when you are working on a local branch not already pushed. If you are working on a branch that exists in the remote, `rebase` can result in some nasty issues so you should prefer a regular `merge`. – Justus Romijn Dec 01 '14 at 08:39
git pull --rebase is fine if you are working with exactly one remote, so your changes are either local only (in which case they are fine to rebase) or are already one the remote (in which case they won't be rebased). The time when git pull --rebase becomes risky is when you are working with multiple remotes. – plugwash Jan 19 '20 at 14:57

score 163 · Answer 12 · edited Dec 24 '20 at 18:17

OK, here is some information about git pull and git fetch, so you can understand the actual differences... in few simple words, fetch gets the latest data, but not the code changes and not going to mess with your current local branch code, but pull get the code changes and merge it your local branch, read on to get more details about each:

git fetch

It will download all refs and objects and any new branches to your local Repository...

Fetch branches and/or tags (collectively, "refs") from one or more other repositories, along with the objects necessary to complete their histories. Remote-tracking branches are updated (see the description of below for ways to control this behavior).

By default, any tag that points into the histories being fetched is also fetched; the effect is to fetch tags that point at branches that you are interested in. This default behavior can be changed by using the --tags or --no-tags options or by configuring remote..tagOpt. By using a refspec that fetches tags explicitly, you can fetch tags that do not point into branches you are interested in as well.

git fetch can fetch from either a single named repository or URL or from several repositories at once if is given and there is a remotes. entry in the configuration file. (See git-config1).

When no remote is specified, by default the origin remote will be used, unless there’s an upstream branch configured for the current branch.

The names of refs that are fetched, together with the object names they point at, are written to .git/FETCH_HEAD. This information may be used by scripts or other git commands, such as git-pull.

git pull

It will apply the changes from remote to the current branch in local...

Incorporates changes from a remote repository into the current branch. In its default mode, git pull is shorthand for git fetch followed by git merge FETCH_HEAD.

More precisely, git pull runs git fetch with the given parameters and calls git merge to merge the retrieved branch heads into the current branch. With --rebase, it runs git rebase instead of git merge.

should be the name of a remote repository as passed to git-fetch1. can name an arbitrary remote ref (for example, the name of a tag) or even a collection of refs with corresponding remote-tracking branches (e.g., refs/heads/:refs/remotes/origin/), but usually it is the name of a branch in the remote repository.

Default values for and are read from the "remote" and "merge" configuration for the current branch as set by git-branch --track.

I also create the visual below to show you how git fetch and git pull working together...

If you like the image then take a look at the git cheat sheet, which is the same sort of things for all git commands... http://ndpsoftware.com/git-cheatsheet.html — Tom, May 20 '18 at 16:42
Doesn't clone also affect the local repository (copying all the history from remote)? — Tom Loredo, Jul 04 '18 at 04:38

th3sly · Answer 13 · 2016-12-09T10:02:01.063

145

enter image description here

This interactive graphical representation is very helpful in understanging git: http://ndpsoftware.com/git-cheatsheet.html

git fetch just "downloads" the changes from the remote to your local repository. git pull downloads the changes and merges them into your current branch. "In its default mode, git pull is shorthand for git fetch followed by git merge FETCH_HEAD."

edited Dec 09 '16 at 10:02

answered Feb 06 '15 at 11:48

th3sly

1,711
1
15
19

20

People, click on the link to interact with the different columns. This cheatsheet is the best resource I've seen to fully understand the differences between each command. – M. Luisa Carrión Oct 04 '16 at 00:41
THis Answer must go to the top – Trect Oct 26 '19 at 06:12
"A picture is worth 1000 words". This webpage is an excellent resource for anyone who wants a simple, pictorial description of Git's various workflows. Thanks for posting this link. – Jim Fischer Aug 30 '20 at 03:59

score 132 · Answer 14 · edited Jun 20 '20 at 09:12

132

Bonus:

In speaking of pull & fetch in the above answers, I would like to share an interesting trick,

`git pull --rebase`

This above command is the most useful command in my git life which saved a lots of time.

Before pushing your new commits to server, try this command and it will automatically sync latest server changes (with a fetch + merge) and will place your commit at the top in git log. No need to worry about manual pull/merge.

Find details at: http://gitolite.com/git-pull--rebase

edited Jun 20 '20 at 09:12

Community

1
1

answered Dec 23 '15 at 15:31

Sazzad Hissain Khan

29,428
20
134
192

4

Nice tip, though it's worth mentioning to new git users that rebase modifies commit hashes (I found that surprising coming from subversion). – AlexMA Sep 20 '16 at 12:57
1

Can you explain what is difference between `git pull` and `git pull --rebase` ? – Shaiju T Jan 17 '18 at 07:02
2

See the stark warning about this method in an answer above: https://stackoverflow.com/a/6011169/241244 – Oct 24 '18 at 08:00

Justus Romijn · Answer 15 · 2014-07-25T07:01:35.107

121

I like to have some visual representation of the situation to grasp these things. Maybe other developers would like to see it too, so here's my addition. I'm not totally sure that it all is correct, so please comment if you find any mistakes.

                                         LOCAL SYSTEM
                  . =====================================================    
================= . =================  ===================  =============
REMOTE REPOSITORY . REMOTE REPOSITORY  LOCAL REPOSITORY     WORKING COPY
(ORIGIN)          . (CACHED)           
for example,      . mirror of the      
a github repo.    . remote repo
Can also be       .
multiple repo's   .
                  .
                  .
FETCH  *------------------>*
Your local cache of the remote is updated with the origin (or multiple
external sources, that is git's distributed nature)
                  .
PULL   *-------------------------------------------------------->*
changes are merged directly into your local copy. when conflicts occur, 
you are asked for decisions.
                  .
COMMIT            .                             *<---------------*
When coming from, for example, subversion, you might think that a commit
will update the origin. In git, a commit is only done to your local repo.
                  .
PUSH   *<---------------------------------------*
Synchronizes your changes back into the origin.

Some major advantages for having a fetched mirror of the remote are:

Performance (scroll through all commits and messages without trying to squeeze it through the network)
Feedback about the state of your local repo (for example, I use Atlassian's SourceTree, which will give me a bulb indicating if I'm commits ahead or behind compared to the origin. This information can be updated with a GIT FETCH).

edited Jul 25 '14 at 07:01

answered Feb 19 '14 at 21:18

Justus Romijn

13,863
4
47
59

Doesn't a `git pull` also perform a merge, i.e. going all the way to the working copy? – Kamiel Wanrooij Mar 24 '14 at 17:28
Good point, yes it will put all the changes in your working copy, and then you can commit it yourself into the local repo. I will update the visual. – Justus Romijn Mar 25 '14 at 07:50
@JustusRomijn Doesn't a pull also update the local repository? Shouldn't there be an asterisk between the origin and working copy asterisks? – user764754 Jan 08 '15 at 09:34
2

@user764754 When you pull, your working copy get the changes (there can also be some conflicts which you might need to resolve). You still have to commit it into your local repository. – Justus Romijn Jan 12 '15 at 07:17
@JustusRomijn:Thanks for the illustration.It would be great if you can make the diagram more comprehensive by illustrating the effects of operations such as reset ,cherry pick on the repository states. – jith912 Jan 17 '15 at 12:42
@jith912 I think it would be a bit outside of the scope of this question topic, but I might create such one and put it on a blog. If I do, I will add the link here for reference. – Justus Romijn Jan 19 '15 at 08:11
@jith912 in one of the answers this link is provided: http://ndpsoftware.com/git-cheatsheet.html – Justus Romijn Feb 13 '15 at 10:35

score 111 · Answer 16 · edited May 31 '17 at 19:41

I have struggled with this as well. In fact I got here with a google search of exactly the same question. Reading all these answers finally painted a picture in my head and I decided to try to get this down looking at the state of the 2 repositories and 1 sandbox and actions performed over time while watching the version of them. So here is what I came up with. Please correct me if I messed up anywhere.

The three repos with a fetch:

---------------------     -----------------------     -----------------------
- Remote Repo       -     - Remote Repo         -     - Remote Repo         -
-                   -     - gets pushed         -     -                     -
- @ R01             -     - @ R02               -     - @ R02               -
---------------------     -----------------------     -----------------------

---------------------     -----------------------     -----------------------
- Local Repo        -     - Local Repo          -     - Local Repo          -
- pull              -     -                     -     - fetch               -
- @ R01             -     - @ R01               -     - @ R02               -
---------------------     -----------------------     -----------------------

---------------------     -----------------------     -----------------------
- Local Sandbox     -     - Local Sandbox       -     - Local Sandbox       -
- Checkout          -     - new work done       -     -                     -
- @ R01             -     - @ R01+              -     - @R01+               -
---------------------     -----------------------     -----------------------

The three repos with a pull

---------------------     -----------------------     -----------------------
- Remote Repo       -     - Remote Repo         -     - Remote Repo         -
-                   -     - gets pushed         -     -                     -
- @ R01             -     - @ R02               -     - @ R02               -
---------------------     -----------------------     -----------------------

---------------------     -----------------------     -----------------------
- Local Repo        -     - Local Repo          -     - Local Repo          -
- pull              -     -                     -     - pull                -
- @ R01             -     - @ R01               -     - @ R02               -
---------------------     -----------------------     -----------------------

---------------------     -----------------------     -----------------------
- Local Sandbox     -     - Local Sandbox       -     - Local Sandbox       -
- Checkout          -     - new work done       -     - merged with R02     -
- @ R01             -     - @ R01+              -     - @R02+               -
---------------------     -----------------------     -----------------------

This helped me understand why a fetch is pretty important.

Not that hard to read: the boxes represent the status of a repo, that in each row changes in time left-to-right after the reported operation in row 2 of the box. The labels R0n are tags in git, and a tag with a + is yet uncommited stuff. Sanbox is used for your working folder, which is different from the repo folder, where commited stuff is stored. — user1708042, Aug 30 '17 at 07:57

score 111 · Answer 17 · edited Jul 14 '20 at 20:45

The Difference between GIT Fetch and GIT Pull can be explained with the following scenario: (Keeping in mind that pictures speak louder than words!, I have provided pictorial representation)

Let's take an example that you are working on a project with your team members. So there will be one main Branch of the project and all the contributors must fork it to their own local repository and then work on this local branch to modify/Add modules then push back to the main branch.

So, Initial State of the two Branches when you forked the main project on your local repository will be like this- (A, B and C are Modules already completed of the project)

Now, you have started working on the new module (suppose D) and when you have completed the D module you want to push it to the main branch, But meanwhile what happens is that one of your teammates has developed new Module E, F and modified C.
So now what has happened is that your local repository is lacking behind the original progress of the project and thus pushing of your changes to the main branch can lead to conflict and may cause your Module D to malfunction.

To avoid such issues and to work parallel with the original progress of the project there are Two ways:

1. Git Fetch- This will Download all the changes that have been made to the origin/main branch project which are not present in your local branch. And will wait for the Git Merge command to apply the changes that have been fetched to your Repository or branch.

So now You can carefully monitor the files before merging it to your repository. And you can also modify D if required because of Modified C.

2. Git Pull- This will update your local branch with the origin/main branch i.e. actually what it does is a combination of Git Fetch and Git merge one after another. But this may Cause Conflicts to occur, so it’s recommended to use Git Pull with a clean copy.

if you could change 'Main Branch' to 'Remote Repo', it would be a great answer. — Fuevo, Jun 04 '18 at 00:17

Selvamani · Answer 18 · 2013-07-17T13:35:56.810

90

We simply say:

git pull == git fetch + git merge

If you run git pull, you do not need to merge the data to local. If you run git fetch, it means you must run git merge for getting the latest code to your local machine. Otherwise, the local machine code would not be changed without merge.

So in the Git Gui, when you do fetch, you have to merge the data. Fetch itself won't make the code changes at your local. You can check that when you update the code by fetching once fetch and see; the code it won't change. Then you merge... You will see the changed code.

edited Jul 17 '13 at 13:35

answered Feb 21 '13 at 13:25

Selvamani

6,590
3
27
41

3

I'd rather say `git pull == git fetch + git merge` :) – melvynkim Jun 07 '13 at 10:38
2

But `git pull --rebase = git fetch + git rebase` – Tino Jul 17 '13 at 07:06

Michael Durrant · Answer 19 · 2016-06-01T00:59:49.483

git fetch pulls down the code from the remote server to your tracking branches in your local repository. If your remote is named origin (the default) then these branches will be within origin/, for example origin/master, origin/mybranch-123, etc. These are not your current branches, they are local copies of those branches from the server.

git pull does a git fetch but then also merges the code from the tracking branch into your current local version of that branch. If you're not ready for that changes yet, just git fetch first.

score 79 · Answer 20 · edited Mar 09 '13 at 20:31

79

git fetch will retrieve remote branches so that you can git diff or git merge them with the current branch. git pull will run fetch on the remote brach tracked by the current branch and then merge the result. You can use git fetch to see if there are any updates to the remote branch without necessary merging them with your local branch.

edited Mar 09 '13 at 20:31

Peter Mortensen

28,342
21
95
123

answered Nov 26 '12 at 21:58

ntanase

1,039
8
6

Pinkesh Sharma · Answer 21 · 2017-06-16T18:15:10.210

Git Fetch

You download changes to your local branch from origin through fetch. Fetch asks the remote repo for all commits that others have made but you don't have on your local repo. Fetch downloads these commits and adds them to the local repository.

Git Merge

You can apply changes downloaded through fetch using the merge command. Merge will take the commits retrieved from fetch and try to add them to your local branch. The merge will keep the commit history of your local changes so that when you share your branch with push, Git will know how others can merge your changes.

Git Pull

Fetch and merge run together often enough that a command that combines the two, pull, was created. Pull does a fetch and then a merge to add the downloaded commits into your local branch.

Honey · Answer 22 · 2021-03-28T12:13:15.063

In simple terms, if you were about to hop onto a plane without any Internet connection...before departing you could just do git fetch origin <branch>. It would fetch all the changes into your computer, but keep it separate from your local development/workspace.

On the plane, you could make changes to your local workspace and then merge it with what you've previously fetched and then resolve potential merge conflicts all without a connection to the Internet. And unless someone had made new changes to the remote repository then once you arrive at the destination you would do git push origin <branch> and go get your coffee.

From this awesome Atlassian tutorial:

The git fetch command downloads commits, files, and refs from a remote repository into your local repository.

Fetching is what you do when you want to see what everybody else has been working on. It’s similar to SVN update in that it lets you see how the central history has progressed, but it doesn’t force you to actually merge the changes into your repository. Git isolates fetched content as a from existing local content, it has absolutely no effect on your local development work. Fetched content has to be explicitly checked out using the git checkout command. This makes fetching a safe way to review commits before integrating them with your local repository.

When downloading content from a remote repository, git pull and git fetch commands are available to accomplish the task. You can consider git fetch the 'safe' version of the two commands. It will download the remote content, but not update your local repository's working state, leaving your current work intact. git pull is the more aggressive alternative, it will download the remote content for the active local branch and immediately execute git merge to create a merge commit for the new remote content. If you have pending changes in progress this will cause conflicts and kickoff the merge conflict resolution flow.

With git pull:

You don't get any isolation.
It doesn't need to be explicitly checked out. Because it implicitly does a git merge.
The merging step will affect your local development and may cause conflicts
It's basically NOT safe. It's aggressive.
Unlike git fetch where it only affects your .git/refs/remotes, git pull will affect both your .git/refs/remotes and .git/refs/heads/

Hmmm...so if I'm not updating the working copy with git fetch, then where am I making changes? Where does Git fetch store the new commits?

Great question. First and foremost, the heads or remotes don't store the new commits. They just have pointers to commits. So with git fetch you download the latest git objects (blob, tree, commits. To fully understand the objects watch this video on git internals), but only update your remotes pointer to point to the latest commit of that branch. It's still isolated from your working copy, because your branch's pointer in the heads directory hasn't updated. It will only update upon a merge/pull. But again where? Let's find out.

In your project directory (i.e., where you do your git commands) do:

ls. This will show the files & directories. Nothing cool, I know.
Now do ls -a. This will show dot files, i.e., files beginning with . You will then be able to see a directory named: .git.
Do cd .git. This will obviously change your directory.
Now comes the fun part; do ls. You will see a list of directories. We're looking for refs. Do cd refs.
It's interesting to see what's inside all directories, but let's focus on two of them. heads and remotes. Use cd to check inside them too.
Any git fetch that you do will update the pointer in the /.git/refs/remotes directory. It won't update anything in the /.git/refs/heads directory.
Any git pull will first do the git fetch, update items in the /.git/refs/remotes directory, then merge with your local and then change the head inside the /.git/refs/heads directory.

A very good related answer can also be found in Where does 'git fetch' place itself?.

Also, look for "Slash notation" from the Git branch naming conventions post. It helps you better understand how Git places things in different directories.

To see the actual difference

Just do:

git fetch origin master
git checkout master

If the remote master was updated you'll get a message like this:

Your branch is behind 'origin/master' by 2 commits, and can be fast-forwarded.
  (use "git pull" to update your local branch)

If you didn't fetch and just did git checkout master then your local git wouldn't know that there are 2 commits added. And it would just say:

Already on 'master'
Your branch is up to date with 'origin/master'.

But that's outdated and incorrect. It's because git will give you feedback solely based on what it knows. It's oblivious to new commits that it hasn't pulled down yet...

Is there any way to see the new changes made in remote while working on the branch locally?

Some IDEs (e.g. Xcode) are super smart and use the result of a git fetch and can annotate the lines of code that have been changed in remote branch of your current working branch. If that line has been changed by both local changes and remote branch, then that line gets annotated with red. This isn't a merge conflict. It's a potential merge conflict. It's a headsup that you can use to resolve the future merge conflict before doing git pull from the remote branch.

Fun tip:

If you fetched a remote branch e.g. did:

git fetch origin feature/123

Then this would go into your remotes directory. It's still not available to your local directory. However, it simplifies your checkout to that remote branch by DWIM (Do what I mean):

git checkout feature/123

you no longer need to do:

git checkout -b feature/123 origin/feature/123

For more on that read here

score 51 · Answer 23 · edited Jun 05 '13 at 15:47

51

The only difference between git pull and git fetch is that :

git pull pulls from a remote branch and merges it.

git fetch only fetches from the remote branch but it does not merge

i.e. git pull = git fetch + git merge ...

edited Jun 05 '13 at 15:47

bcr

1,626
21
25

answered Jun 04 '13 at 14:08

Rohitashv Singhal

4,261
13
48
100

1

And neither help if git thinks you are behind by commits and can "fast-forward", which upon I ended up `rm -rf`ing the whole thing and starting over. Stupid Git, please just let me get current so I can go back to work? – Chris K Sep 11 '13 at 22:01

score 44 · Answer 24 · answered Nov 28 '13 at 17:03

Git allows chronologically older commits to be applied after newer commits. Because of this, the act of transferring commits between repositories is split into two steps:

Copying new commits from remote branch to copy of this remote branch inside local repo.

(repo to repo operation) master@remote >> remote/origin/master@local
Integrating new commits to local branch

(inside-repo operation) remote/origin/master@local >> master@local

There are two ways of doing step 2. You can:

Fork local branch after last common ancestor and add new commits parallel to commits which are unique to local repository, finalized by merging commit, closing the fork.
Insert new commits after last common ancestor and reapply commits unique to local repository.

In git terminology, step 1 is git fetch, step 2 is git merge or git rebase

git pull is git fetch and git merge

score 37 · Answer 25 · answered Aug 12 '14 at 04:00

Git obtains the branch of the latest version from the remote to the local using two commands:

git fetch: Git is going to get the latest version from remote to local, but it do not automatically merge. git fetch origin master git log -p master..origin/master git merge origin/master

The commands above mean that download latest version of the main branch from origin from the remote to origin master branch. And then compares the local master branch and origin master branch. Finally, merge.
git pull: Git is going to get the latest version from the remote and merge into the local.

git pull origin master

The command above is the equivalent to git fetch and git merge. In practice, git fetch maybe more secure because before the merge we can see the changes and decide whether to merge.

Donal · Answer 26 · 2015-07-11T10:22:46.543

What is the difference between git pull and git fetch?

To understand this, you first need to understand that your local git maintains not only your local repository, but it also maintains a local copy of the remote repository.

git fetch brings your local copy of the remote repository up to date. For example, if your remote repository is GitHub - you may want to fetch any changes made in the remote repository to your local copy of it the remote repository. This will allow you to perform operations such as compare or merge.

git pull on the other hand will bring down the changes in the remote repository to where you keep your own code. Typically, git pull will do a git fetch first to bring the local copy of the remote repository up to date, and then it will merge the changes into your own code repository and possibly your working copy.

score 35 · Answer 27 · answered Sep 19 '13 at 22:41

git pull == ( git fetch + git merge)

git fetch does not changes to local branches.

If you already have a local repository with a remote set up for the desired project, you can grab all branches and tags for the existing remote using git fetch . ... Fetch does not make any changes to local branches, so you will need to merge a remote branch with a paired local branch to incorporate newly fetch changes. from github

score 35 · Answer 28 · answered Jul 29 '15 at 18:38

Trying to be clear and simple.

The git pull command is actually a shortcut for git fetch followed by the git merge or the git rebase command depending on your configuration. You can configure your Git repository so that git pull is a fetch followed by a rebase.

score 35 · Answer 29 · answered Jul 12 '17 at 06:23

A simple Graphical Representation for Beginners,

here,

git pull

will fetch code from repository and rebase with your local... in git pull there is possibility of new commits getting created.

but in ,

git fetch

will fetch code from repository and we need to rebase it manually by using git rebase

eg: i am going to fetch from server master and rebase it in my local master.

1) git pull ( rebase will done automatically):

git pull origin master

here origin is your remote repo master is your branch

2) git fetch (need to rebase manually):

git fetch origin master

it will fetch server changes from origin. and it will be in your local until you rebase it on your own. we need to fix conflicts manually by checking codes.

git rebase origin/master

this will rebase code into local. before that ensure you're in right branch.

Nice graph, but you might want to explain why you use "rebase" when the graph says "merge". — Guntram Blohm, Oct 10 '18 at 09:33
merge will represent another branch commit and produce new commit which contains commits as a reference. but rebase will replicate commits from another branch it wont create new commit rather than it replicate — Mohideen bin Mohammed, Oct 10 '18 at 11:00

score 34 · Answer 30 · edited Aug 17 '17 at 20:51

Actually Git maintains a copy of your own code and the remote repository.

The command git fetch makes your local copy up to date by getting data from remote repository. The reason we need this is because somebody else might have made some changes to the code and you want to keep yourself updated.

The command git pull brings the changes in the remote repository to where you keep your own code. Normally, git pull does this by doing a ‘git fetch’ first to bring the local copy of the remote repository up to date, and then it merges the changes into your own code repository and possibly your working copy.

score 31 · Answer 31 · answered Jul 12 '15 at 03:54

31

git pull = git fetch + git merge

answered Jul 12 '15 at 03:54

Saqib R.

2,559
2
24
19

score 25 · Answer 32 · edited May 23 '17 at 12:26

git pull

It performs two functions using a single command.

It fetches all the changes that were made to the remote branch and then merges those changes into your local branch. You can also modify the behaviour of pull by passing --rebase. The difference between merge and rebase can be read here

git fetch

Git fetch does only half the work of git pull. It just brings the remote changes into your local repo but does not apply them onto your branches.You have to explicitly apply those changes. This can be done as follows:

git fetch
git rebase origin/master

score 23 · Answer 33 · answered Sep 25 '15 at 09:46

One must keep in mind the nature of git. You have remotes and your local branches ( not necessarily the same ) . In comparison to other source control systems this can be a bit perplexing.

Usually when you checkout a remote a local copy is created that tracks the remote.

git fetch will work with the remote branch and update your information.

It is actually the case if other SWEs are working one the same branch, and rarely the case in small one dev - one branch - one project scenarios.

Your work on the local branch is still intact. In order to bring the changes to your local branch you have to merge/rebase the changes from the remote branch.

git pull does exactly these two steps ( i.e. --rebase to rebase instead of merge )

If your local history and the remote history have conflicts the you will be forced to do the merge during a git push to publish your changes.

Thus it really depends on the nature of your work environment and experience what to use.

Sujeet Agrahari · Answer 34 · 2020-05-25T05:55:50.407

All branches are stored in .git/refs

All local branches are stored in .git/refs/heads

All remote branches are stored in .git/refs/remotes

The git fetch command downloads commits, files, and refs from a remote repository into your local repo. Fetching is what you do when you want to see what everybody else has been working on.

So when you do git fetch all the files, commits, and refs are downloaded in

this directory .git/refs/remotes

You can switch to these branches to see the changes.

Also, you can merge them if you want.

git pull just downloads these changes and also merges them to the current branch.

Example

If you want see the work of remote branch dev/jd/feature/auth, you just need to do

git fetch origin dev/jd/feature/auth

to see the changes or work progress do,

git checkout dev/jd/feature/auth

But, If you also want to fetch and merge them in the current branch do,

git pull origin dev/jd/feature/auth

If you do git fetch origin branch_name, it will fetch the branch, now you can switch to this branch you want and see the changes. Your local master or other local branches won't be affected. But git pull origin branch_name will fetch the branch and will also merge to the current branch.

score 8 · Answer 35 · answered May 13 '20 at 16:32

8

Simple explanation:

git fetch

fetches the metadata. If you want to check out a recently created branch you may want to do a fetch before checkout.

git pull

Fetches the metadata from remote and also moved the files from remote and merge on to the branch

answered May 13 '20 at 16:32

Akash Yellappa

1,186
13
16

2

If you decide to answer an older question that has well established and correct answers, adding a new answer late in the day may not get you any credit. If you have some distinctive new information, or you're convinced the other answers are all wrong, by all means add a new answer, but 'yet another answer' giving the same basic information a long time after the question was asked usually won't earn you much credit. – Jonathan Leffler May 19 '20 at 05:13
2

I think this is one of the better answers, though it's more than just metadata. I would say that "git fetch only modifies files in the hidden .git folder, including metadata and copies of all the remote branches, whereas git pull also changes the files in your current local branch". If someone had explained it that way a few years ago my transition from svn to git would have been a lot faster! All these other answers have a ridiculous amount of exposition that is not particularly helpful IMO. – WiringHarness Apr 12 '21 at 23:13
Glad you feel so – Akash Yellappa Apr 14 '21 at 20:39

score 5 · Answer 36 · answered Dec 22 '17 at 20:36

Git Fetch

Helps you to get known about the latest updates from a git repository. Let's say you working in a team using GitFlow, where team working on multiple branches ( features ). With git fetch --all command you can get known about all new branches within repository.

Mostly git fetch is used with git reset. For example you want to revert all your local changes to the current repository state.

git fetch --all // get known about latest updates
git reset --hard origin/[branch] // revert to current branch state

Git pull

This command update your branch with current repository branch state. Let's continue with GitFlow. Multiple feature branches was merged to develop branch and when you want to develop new features for the project you must go to the develop branch and do a git pull to get the current state of develop branch

Documentation for GitFlow https://gist.github.com/peterdeweese/4251497

score 4 · Answer 37 · edited Dec 03 '20 at 07:27

4

This graphic could be of help. git pull is essentially equivalent to git fetch then git merge

edited Dec 03 '20 at 07:27

Steven

904
1
9
22

answered Nov 19 '20 at 09:20

Piaget Hadzizi

408
2
10

git merge does not only affect your working copy... – masterxilo Dec 15 '20 at 19:34

score 0 · Answer 38 · answered Dec 15 '20 at 19:32

A git repository contains immutable data blobs and a few mutable pointers/references/names (we call them branches, HEADS) for usability (in theory, git could be a pure append-only store, accessed only by commit hashes).

The immutable blobs are usually always eventually shared among all contributors. Every developer has a copy of all of these on his machine.

git fetch downloads the latest blobs and remote mutable files to your machine.

It does not change your mutable files nor create any blobs that did not previously exist somewhere.

git pull is git fetch then git merge. git merge creates new blobs that never existed in the remote repository before and updates your mutable files (your references).

What is the difference between 'git pull' and 'git fetch'?

38 Answers38

UPDATE

git fetch

git pull

Bonus:

`git pull --rebase`

To see the actual difference

Is there any way to see the new changes made in remote while working on the branch locally?

Fun tip:

Linked

Related