How to Compare Files one by one using Shell Script

[Padma Asrani]

Hello All

I have two directories and I wanted to compare files one by one in these directories using shell script. Can anybody help how should I start looping through files in each directory and compare these line by line and get the difference?

Thanks
Padma

[OldSchool]

2 directories, compare files.....

you're going to have to be a little more specific...

Compare file names only?
Compare like named files?
Text only, or are there binaries, such as executables?
Desired output?

"man diff".....

[Padma Asrani]

Hello

These are the text files and I mentioned that I wanted to compare these files line by line.

Regards
Padma

[James R. Ferguson]

Hi Padma:

Have a look at the manpages for 'diff' and for 'dircmp':

http://docs.hp.com/en/B3921-60631/diff.1.html

http://docs.hp.com/en/B3921-60631/dircmp.1.html

Regards!

...JRF...

[TwoProc]

Since you've not given us much to go on, I'll have to just make assumptions.

Assumptions:
1) both directories have all the same number of files, only their content may be different.
2) both directories are siblings of the same parent directory.
3) There are no subdirectories that you want to scan.
4) There are no missing files that you are, or are not trying to compare.
5) The two sibling directories are named "a" and "b". And they are children of the current directory "."

$> cd a
for i in *
do
diff $i ../b/$i
done

[Peter Nikitka]

Hi,

a call of
dircmp -d -s dira dirb

will
- no identical file
- flag files only present in one of the directories
- list a diff-report of identical filenames in both directories

mfG Peter

[TwoProc]

Wow,

I forgot all about dircmp.

> Do what they say, use dircmp.

[Padma Asrani]

ok. People thanks a lot. I am trying all the options. Thanks so much to all.

[Padma Asrani]

Hi

Some more clarification. Both the directories have some files common and also files which are in one directory but not in other. How to get the list of uncommon files?

Secondly I am not sure i m not able to find dircmp on my linux box.

Thanks
Padma

[Pete Randall]

From the dircmp man page:

WARNINGS
This command is likely to be withdrawn from X/Open standards. Applications using this command might not be portable to other vendors' systems. As an alternative diff -R is recommended.

You might want to take a look at diff -R.


Pete

[James R. Ferguson · Accepted solution]

Hi Padma:

> Both the directories have some files common and also files which are in one directory but not in other. How to get the list of uncommon files?

Using 'diff' would show you this automatically. Consider:

# ls /tmp/dir1 /tmp/dir2

/tmp/dir1:
hosts inetd.conf

/tmp/dir2:
hosts services

# diff /tmp/dir1 /tmp/dir2
17c17
< 127.0.0.1 LOCALHOST loopback
---
> 127.0.0.1 localhost loopback
Only in /tmp/dir1: inetd.conf
Only in /tmp/dir2: services
#

...at worst, you could 'grep -v "^Only in"' to filter out the unmatched files; or 'grep "^Only in"' to find them!

> Secondly I am not sure i m not able to find dircmp on my linux box.

It probably isn't. In fact, the manpages for 'dircmp' note, "This command is likely to be withdrawn from X/Open standards. Applications using this command might not be portable to other vendors' systems. As an alternative diff -R is recommended."

Regards!

...JRF...

[Padma Asrani]

Thanks

it does work.

Regards
Padma

[Padma Asrani]

Hi

I was not able to find diff -R and I just used diff dir2> |grep "^Only in"

diff -R is not supported on my Linux box. I am using Fedora Linux.

Thanks
Padma

[OldSchool]

thanks....

[James R. Ferguson]

Hi Padma:

> diff -R is not supported on my Linux box. I am using Fedora Linux.

No, but:

# diff -r

...and I'm running Fedora 8 in this case.

Your first line of defense is to do:

# man diff

Regards!

...JRF...

[Dennis Handly]

>Both the directories have some files common and also files which are in one directory but not in other. How to get the list of uncommon files?

You can just list both directories and use comm(3) to get this:
$ ls dir1 > dir_files.1
$ ls dir2 > dir_files.2
# Gives files only in dir1
$ comm -23 dir_files.1 dir_files.2
# Gives files only in dir2
$ comm -13 dir_files.1 dir_files.2
# Gives files that aren't in both:
# Ones in dir2 are indented
$ comm -3 dir_files.1 dir_files.2

[TwoProc]

Just an fyi -

I'm guessing you've realized it by now - but you're getting "HPUX" answers because you've posted in the HPUX forum. There is a Linux forum, in which you're much more likely to get Linux specific help.

[James R. Ferguson]

Hi (again):

> but you're getting "HPUX" answers because you've posted in the HPUX forum

There was never a specification of what operating system but I think we have covered the options available to (L)UNIX at large.

Regards!

...JRF...

[Rasheed Tamton]

Hi Padma,

This one I mostly use:

for i in `find /tmp/dir1 -type f|awk -F/ '{print $NF}'`
do
find /tmp/dir2 -type f -exec ll {} \; |grep $i
done

Can be ported to any UNIX.

Regards,
Rasheed Tamton.