There are dozens of ways I imagine. I'd probably choose a lookahead: $line =~ s/\s+(?=\|)//g. Hrm, I just noticed there's space after one of them. A look-ahead is probably the wrong choice then. I'd probably choose $line =~ s/\s*\|\s*/\|/g instead — there may be a more efficient way though.

I won't guarantee it's use in a production environment though. I've never tested the above.

And in that vein, the following command should scan through a file and perform the operation on every line, leaving a backup with .bak appended. If this is Windows, change the single quotes to double quotes.

perl -pi.bak -e 's/\s*\|\s*/\|/g' filename

You might try something like this:

#!/usr/bin/perl -w
use strict;
use warnings;

my $in = <DATA>;
my @fields =
        # Remove leading & trailing space from each list element
        map { s/^\s+//; s/\s+$//; $_ }
        # Split input string into fields at '|'
        split /\|/, $in;

print join('|',@fields), "\n";
__DATA__
abcd |abcd|abcd   |      abcd       | abcd|
[download]