Question

Looking for an nice (short & elegant) one liner for the next perl script:

use strict;
use warnings;

my %all;
while(<DATA>) {
    chomp;
    my ($name, $x, $path) = split /\s+/;
    push @{$all{$path}}, $name;
}
foreach my $path (sort keys %all) {
    my $cnt = scalar @{$all{$path}};
    print "$path $cnt @{$all{$path}}\n" if $cnt > 1;
}
__DATA__
Atxt  x     a/b/c
Btxt x a/d/x
Ctxt x i/t/a
Dtxt x i/y/a
Etxt  x     i/t/a
Ftxt x a/d/x
Gtxt x a/d/x

ofc, the one-liner should read from STDIN not from DATA.

In short, the script read 3 fields (name, x, path), and should output summary for duplicated paths in a form: path dup_count name1 ... namex. Each name is different.

So, looking for something like:

my_command | perl -F '\s+' -nle 'shorter_variant_of_the_above_script'

Answer 1

There are probably ways to shorten this, but here’s an attempt:

perl -lane 'push@{$h{$F[2]}},$F[0]}{map{$n=@v=@{$h{$_}};print"$_ $n @v"if$n>1}keys%h'

Output with provided data:

a/d/x 3 Btxt Gtxt Ftxt
i/t/a 2 Etxt Ctxt