Getting exact match in a has but with a twist - arrays

Question

I have something I cannot get my head around
Let's say I have a phone list used for receiving and dialing out stored like below. The from and to location is specified as well.
Country1 Country2 number1 number2
USA_Chicago USA_LA 12 14
AUS_Sydney USA_Chicago 19 15
AUS_Sydney USA_Chicago 22 21
CHI_Hong-Kong RSA_Joburg 72 23
USA_LA USA_Chigaco 93 27
Now all I want to do is to remove all the duplicates and give only what is relevant to the countries as keys and each number that is assigned to it in a pair, but the pair needs to be bi-directional.
In other words I need to get results back and then print them like this.
USA_Chicago-USA_LA 27 93 12 14
Aus_Sydney-USA_Chicago 19 15 22 21
CHI_Hong-kong-RSA_Joburg 72 23
I have tried many methods including a normal hash table and the results seem fine, but it does not do the bi-direction, so I will get this instead.
USA_Chicago-USA_LA 12 14
Aus_Sydney-USA_Chicago 19 15 22 21
CHI_Hong-kong-RSA_Joburg 72 23
USA_LA-USA_Chicago 93 27
So the duplicate removal works in one way, but because there is another direction, it will not remove the duplicate "USA_LA-USA_Chicago" which already exists as "USA_Chicago-USA_LA" and will store the same numbers under a swopped name.
The hash table I tried last is something like this. (not exactly as I trashed the lot and had to rewrite it for this post)
#input= ("USA_Chicago USA_LA 12 14" ,
"AUS_Sydney USA_Chicago 19 15" ,
"AUS_Sydney USA_Chicago 22 21" ,
"CHI_Hong-Kong RSA_Joburg 72 23" '
"USA_LA USA_Chigaco 93 27");
my %hash;
for my $line (#input) {
my ($c1, $c2, $n1, $n2) = split / [\s\|]+ /x, $line6;
my $arr = $hash{$c1} ||= [];
push #$arr, "$n1 $n2";
}
for my $c1 (sort keys %hash) {
my $arr = $hash{$c1};
my $vals = join " : ", #$arr;
print "$c1 $vals\n";
}
So all if A-B exists and so does B-A, use only one but assign the values from the key being removed, to the remaining key. I basically need to do is get rid of any duplicate key in any direction, but assign the values for to the remaining key. So A-B and B-A would be considered a duplicate, but A-C and B-C are not. -_-

Answer 1

Simply normalise the destinations. I chose to sort them.
use strictures;
use Hash::MultiKey qw();
my #input = (
'USA_Chicago USA_LA 12 14',
'AUS_Sydney USA_Chicago 19 15',
'AUS_Sydney USA_Chicago 22 21',
'CHI_Hong-Kong RSA_Joburg 72 23',
'USA_LA USA_Chicago 93 27'
);
tie my %hash, 'Hash::MultiKey';
for my $line (#input) {
my ($c1, $c2, $n1, $n2) = split / [\s\|]+ /x, $line;
my %map = ($c1 => $n1, $c2 => $n2);
push #{ $hash{[sort keys %map]} }, #map{sort keys %map};
}
__END__
(
['CHI_Hong-Kong', 'RSA_Joburg'] => [72, 23],
['AUS_Sydney', 'USA_Chicago'] => [19, 15, 22, 21],
['USA_Chicago', 'USA_LA'] => [12, 14, 27, 93],
)

Answer 2

Perl is great for creating complex data structures but learning to use them effectively takes practices.
Try:
#!/usr/bin/env perl
use strict;
use warnings;
# --------------------------------------
use charnames qw( :full :short );
use English qw( -no_match_vars ); # Avoids regex performance penalty
use Data::Dumper;
# Make Data::Dumper pretty
$Data::Dumper::Sortkeys = 1;
$Data::Dumper::Indent = 1;
# Set maximum depth for Data::Dumper, zero means unlimited
local $Data::Dumper::Maxdepth = 0;
# conditional compile DEBUGging statements
# See http://lookatperl.blogspot.ca/2013/07/a-look-at-conditional-compiling-of.html
use constant DEBUG => $ENV{DEBUG};
# --------------------------------------
# skip the column headers
<DATA>;
my %bidirectional = ();
while( my $line = <DATA> ){
chomp $line;
my ( $country1, $country2, $number1, $number2 ) = split ' ', $line;
push #{ $bidirectional{ $country1 }{ $country2 } }, [ $number1, $number2 ];
push #{ $bidirectional{ $country2 }{ $country1 } }, [ $number1, $number2 ];
}
print Dumper \%bidirectional;
__DATA__
Country1 Country2 number1 number2
USA_Chicago USA_LA 12 14
AUS_Sydney USA_Chicago 19 15
AUS_Sydney USA_Chicago 22 21
CHI_Hong-Kong RSA_Joburg 72 23
USA_LA USA_Chicago 93 27