[SOLVED] How to remove duplicate files from Two Folders?

guessity · 08-19-2010, 12:26 AM

I have two folders -

Folder abc and Folder xyz which contains 1000's of files with few of them having the same file names.

How can I remove the duplicates from Folder abc?

plz help me out on this.

grail · 08-19-2010, 12:53 AM

Are there other folders in these folders?
Assuming not you could try a simple for loop:

Code:

for file in xyz/*
do
    [[ -e "abc/$file" ]] && echo "abc/$file"
done

Change echo to rm if the data looks correct. If there are spaces in files names may need to choose an alternate while loop combo.

guessity · 08-19-2010, 01:04 AM

Hi Grail,
thx. I did the following

Code:

#!/bin/bash
for file in /home/tst/xyz/*
do
    [[ -e "/home/tst/abc/$file" ]] && rm -rf "/home/tst/abc/$file"
done

However I can see nothing coming. I have a file in the both the folders called 1.txt and that is not getting removed for some reason.

Any idea where I have gone wrong?

quanta · 08-19-2010, 01:22 AM

Try fdupes.

guessity · 08-19-2010, 01:26 AM

Actually I have two directories and I dont have permission to install on this linux system.

EricTRA · 08-19-2010, 01:28 AM

Quote:

Originally Posted by guessity

Hi Grail,
thx. I did the following

Code:

#!/bin/bash
for file in /home/tst/xyz/*
do
    [[ -e "/home/tst/abc/$file" ]] && rm -rf "/home/tst/abc/$file"
done

However I can see nothing coming. I have a file in the both the folders called 1.txt and that is not getting removed for some reason.

Any idea where I have gone wrong?

Hello,

You need to look at your paths and variables. You have set full path so when calling your variable $file you get the full path, so nothing found because:

Code:

$file will be /home/tst/xyz/filename

and in your test:

Code:

[[ -e "/home/tst/abc//home/tst/xyz/filename" ]]

Change your script like this and run it from /home/tst (or put that in another variable)

Code:

#!/bin/bash
for file in xyz/*
do
    filename=`basename $file`
    [[ -e "abc/$filename" ]] && rm -rf "abc/$filename"
done

Kind regards,

Eric

EricTRA · 08-19-2010, 01:31 AM

Also, you can 'dry run' your script by putting this line:

Code:

set -x

right after #!/bin/bash. That way the script doesn't get executed but you'll get output on your screen of what would be done so you can see where it goes wrong. Works great when troubleshooting.

Kind regards,

Eric

guessity · 08-19-2010, 01:36 AM

Thx EricTRA. It works well!!

EricTRA · 08-19-2010, 01:43 AM

Hi,

Glad it works.

Kind regards,

Eric

ghostdog74 · 08-19-2010, 01:43 AM

Code:

echo rm -f $(sort <(ls -1 abc) <(ls -1 xyz)| sed 's/^\.*\///' | uniq -d | sed -e ":a" -e 'N;s/\n/ xyz\//;ba')

i92guboj · 08-19-2010, 01:50 AM

You might want to double check that the files are exact copies first, though, otherwise you might lose something important.

Code:

#!/bin/bash
for file in xyz/*
do
    filename=`basename $file`
    if [[ -e "abc/$filename" ]]; then
        if diff -q "$file" "abc/$filename" > /dev/null; then
            rm -rf "abc/$filename"
        else
            echo "$file and abc/$filename are different, review manually."
        fi
    fi
done

grail · 08-19-2010, 04:05 AM

Sorry my response didn't work ... wasn't paying attention. In my defence however, I did suggest using echo before using rm to check data was correct.
This would have shown you the path issue straight away

Leslie007 · 08-19-2010, 04:26 AM

You can use this perl script also,

use strict;
use warnings;
use File::Basename;
my @dir = </home/Bash/abc/*>;

foreach my $file (@dir) {
unlink($file) if(-e "/home/test/xyz/".basename($file)) ;
}

EricTRA · 08-19-2010, 04:28 AM

Hi grail,

Your response does work if you assume executing the script from the 'base' location. Since OP didn't use that structure but put in the full path it didn't work as expected.

Kind regards,

Eric

konsolebox · 08-19-2010, 05:57 AM

I was wondering... abc/* will form abc/file1, abc/file2 .. etc. So how did your scripts work?
Shouldn't it be just like this?

Code:

cd abc/
for A in *; do
    [[ -e ../xyz/$A ]] && rm "$A"
done

I commonly do this pattern when I'm deleting duplicate files in a directory.