How To Compare Directories in Unix

Certain situations require you to quickly confirm which files between two directories are different, and while your particular requirements may suggest writing a script for this task, I want to make sure you're familiar with the basics first – majority of directory comparisons can be done using diff command (yes, that's right – the same one used for comparing files).

Why compare directories?

First of all, let's agree on why you may need to compare directories. There's a few possible reasons:

  • comparing the amount of space consumed by two directories - this is the very first and the fasted way to compare directories because it gives you an idea how close in terms of space usage the directories are. For example, if you're comparing two daily backups of the same piece of software, you normally don't expect them to be vastly different.
  • identifying if some files are missing from one of the directories – can be useful when you want to make sure two directories with configuration files for a certain package are identical – files can be different, but the same files are present in the same locations for both directories
  • confirming if files in two directories are the same – a typical task when comparing your actual data against a backup copy. When something goes wrong, this is one of the first things you do to make sure all the important files are not only present, but are actually the same as they have been when you took the last backup copy
  • highlighting textual differences between files in directories – this is a useful exercise when you're looking at two similar directories and expect only minor changes between the files – version numbers, different file or directory names hardcoded in various scripts, etc.

Comparing the size of two directories

I'm going to show you this trick before getting into details of using diff command. For size comparison, we should use the du command, it's really easy.

The options used for the du command in the example below are: -s for summary (calculate the directory size based on the sizes of all the possible subdirectories it may have) and -k for kilobytes, so /usr/lib is roughly 400Mb in size as per the output below.

ubuntu$ du -sk /usr/lib /usr/lib64
404196  /usr/lib
0       /usr/lib64

This sample output will tell you that directories are vastly different, so that may save you time because you may choose not to compare anything file-by-file if one of the directories looks to be empty or really off space consumption wise.

Test setup for diff comparison exercises

For today's post, I've created a set of directories and files to show how you can compare them. Here is the setup:

ubuntu$ find /tmp/dir1 /tmp/dir2
/tmp/dir1
/tmp/dir1/file1
/tmp/dir1/file2
/tmp/dir1/dir11
/tmp/dir1/dir11/file11
/tmp/dir1/dir11/file12
/tmp/dir2
/tmp/dir2/file1
/tmp/dir2/dir11
/tmp/dir2/dir11/file11
/tmp/dir2/dir11/file12
/tmp/dir2/file3

As you can see, I've got two directories: /tmp/dir1 and /tmp/dir2, with a dir11 subdirectory in each of them. There's also a few files here and there, some of them missing from one of the directories specifically to be highlighted by our comparison exercises.

Basic diff usage for comparing directories

The easiest way to get started is to simply invoke diff command and specify two directories as command line parameters. Here's what you will probably see:

ubuntu$ diff /tmp/dir1 /tmp/dir2
Common subdirectories: /tmp/dir1/dir11 and /tmp/dir2/dir11
diff /tmp/dir1/file1 /tmp/dir2/file11
Only in /tmp/dir1: file2
Only in /tmp/dir2: file3

This output confirms that /tmp/dir1 and /tmp/dir2 both contain a dir11 directory, and also shows that /tm/dir1/file1 and /tmp/dir2/file1 are actually different files even though they have the same name. By default, diff compares such files and you can see the result of each comparison in the output. Also included are pointers to the files which are present only in one of the compared directories: you can see that file2 can only be found in /tmp/dir1 and file3 was present only in /tmp/dir2.

Find which files are missing in one of the directories

From the example below, it is easy to deduct that the command line for identifying files missing in one of the directories will be this one:

ubuntu$ diff /tmp/dir1 /tmp/dir2 | grep Only
Only in /tmp/dir1: file2
Only in /tmp/dir2: file3

Highlight the different files, not the differences

If you're only interested in files which exist in both directory structures, but are different – you can use a special command line option. It will simply point the files out, without getting into any further details. You'll probably notice how this output is very similar to the default one:

ubuntu$ diff --brief /tmp/dir1 /tmp/dir2
Common subdirectories: /tmp/dir1/dir11 and /tmp/dir2/
Files /tmp/dir1/file1 and /tmp/dir2/file1 differ
Only in /tmp/dir1: file2
Only in /tmp/dir2: file3

Note how instead of showing the difference between file1 in /tmp/dir1 and /tmp/dir2, this time you only get told that these two files are different.

How to recursively compare directories

If you're dealing with a complex directory structure, you'll be glad to know that –recursive parameter for the diff command compares not only the immediate directories pointed to from the command line, but also walks through the full tree of subdirectories:

ubuntu$ diff --recursive --brief /tmp/dir1 /tmp/dir2
Files /tmp/dir1/dir11/file12 and /tmp/dir2/dir11/file12 differ
Files /tmp/dir1/file1 and /tmp/dir2/file1 differ
Only in /tmp/dir1: file2
Only in /tmp/dir2: file3

Feeling better now? Many directory comparison tasks can be accomplished using the diff command, but if you're stuck with a particular problem which can't be solved using my examples – please leave a commend and I'll come up with a solution.

See also:

Share and Enjoy

  • Facebook
  • Twitter
  • Delicious
  • LinkedIn
  • StumbleUpon
  • Add to favorites
  • Email
  • RSS
  • rasa

    Hi ,

    I have a case,where i need to copy only the changed/new files from one location to another.Say i have a 2 directories version1 and version2.I need to copy only the new/changed files from version 2 to version1.Basically directories version2 and version1 should be in sync,i should not blindly copy version2 dir recursively into version.Im working on writing a script for this,by checking the diff and if changed copying them….do you have a script for this r any commad to do it

  • Gleb Reys

    Hi Rasa,

    Your task can easily be accomplished using rsync command, look up a man page for it for more info or stay tuned – I'll post a quick tutorial for it soon enough.

  • rasa

    Thanks,
    It seems to answer my problem.
    Does it write any log for the list of files changed/added in the destination.
    say i use rsync -av version2 version1
    I need to find out what are the files changed/added/removed in version1 after this command…Just a log

  • Gleb Reys

    Yes, rasa, rsync provides you with a list of files changed or removed as the result.

  • jack

    hi ,
    i have got one problem due to failure of one hard drive as follows:
    1:i had 3 drives having images data.i was downloading data from all 3 drives and one of 3 drives got failure so i lost all data in broken drive but 2 other drives data are safe(2 drives are safe drives).
    2. i did contact with administrator and he restored the original data with for 6 days data.
    3. Restored data contains data from broken drive and safe drive also.
    4. my problem is to get only data that should be in broken drive but not in safe drives.
    because safe drives data already got transferred.

    please give me solution or any idea how to get only broken drive data from restored .
    i was thinking to write shell script but i m new in this.
    help will be higly appriciated.
    thnx a ton

  • jack

    hi arun,
    i have got one problem due to failure of one hard drive as follows:
    1:i had 3 drives having images data.i was downloading data from all 3 drives and one of 3 drives got failure so i lost all data in broken drive but 2 other drives data are safe(2 drives are safe drives).
    2. i did contact with administrator and he restored the original data with for 6 days data.
    3. Restored data contains data from broken drive and safe drive also.
    4. my problem is to get only data that should be in broken drive but not in safe drives.
    because safe drives data already got transferred.

    please give me solution or any idea how to get only broken drive data from restored .
    i was thinking to write shell script but i m new in this.
    help will be higly appriciated.
    thnx a ton

  • Pingback: How To Syncronize Directories with Rsync | UNIX Tutorial: Learn UNIX

  • Robin

    Is there some way to copy the files (that you figure from the diff command) are different in the two directories into a new directory?

  • Prashant Deshani

    Hi,

    First of all Many Thanks for the nice explaination of comparison of 2 directories.

    However, I had the following concerns when I tried to compare the 2 directories:

    1. I wanted to neglect the "Carriage Return" changes, but it seems like "Diff" command does not come with that functionality.

    2. The command "–brief" does not seem to work.

    3. The command lists the file which are identical…but without any difference. How can we avoid that?

    Following is the snapshot of the ouput:

    diff -bitw -r /mbx/cok/compare1/cokmyep.6rq /mbx/cok/compare2/cokmyep.6rq
    diff -bitw -r /mbx/cok/compare1/cokmyep_2.6rq /mbx/cok/compare2/cokmyep_2.6rq
    1a2,7
    >
    >
    >
    >
    >
    >
    diff -bitw -r /mbx/cok/compare1/cokmyep_3.6rq /mbx/cok/compare2/cokmyep_3.6rq
    Only in /mbx/cok/compare2: comparison.txt
    Only in /mbx/cok/compare2: remove_M.sh

  • Chorroborro

    Is there a way to add the extension .bad to the files that are in one dir but no in the other and then have them listed ina log and then printed into paper with red ink?

  • http://www.unixtutorial.org/2008/06/how-to-compare-directories-in-unix/ sunil

    i want to difference of two directories and show files which are present in both.

    out put shouldnot show the resulting similar directories

  • Pingback: Vantage Point of Queens - 090823: learn

  • Deepak R Kadam

    awesome work… ;)

  • Pingback: links for 2011-03-21 – Kevin Burke

  • andrea

    :D great tyvm!!!!!!!!!!

  • Hanna Barbera

    Life saver!!! i knew of the diff command and i knew it would compare files and such but i didint know how to use it and the man page is confusing (as most are). But you shoudl see some of the crazy crap you get when you google this subject try googling "compare two directories linux" you get some crazy solutions to a simple problem…lol anyways thanx [bookmarked]