use Data::Dumper;
my %fruits=();
my @arr = ([1,1],[1,2],[3,5],[6,7]);
for (@arr)
{
  push @{$fruits{$_->[0]}}, $_->[1];
}

print Dumper(\%fruits);
[download]

use Data::Dumper;
@ar=([1,1],[1,2],[3,5],[6,7]);

map {test($_->[0],$_->[1])} @ar;

sub test
{
    my ($k,$val)=@_;
    
    if (exists $hash{$k})
    {
        $hash{$k} .= ','.$val ;
    }
    else
    {
        $hash{$k}=$val;
    }
}

print Dumper(\%hash);
[download]

Output:

$VAR1 = {
          '6' => 7,
          '1' => '1,2',
          '3' => 5
        };
[download]

There's two possible and different ways to do this.

push @{$fruits{$a}},$b
[download]

will create a hash-of-arrays where each hash element references an array.

$fruits{$a} .= ','.$b ;
[download]

will just concatenate all values($b) for a particular key ($a).
The selection of one of these depends entirely on what is to be done with the hash after it has been populated. For computational purposes, dealing with a hash-of-arrays is (IMHO) easier than doing a series of splits on a value which was constructed by concatenating...
The hash-of-arrays is easier to iterate through, whereas if you just need to print to output, concatenating might be a better option.

Manav

$fruits{$a} = $b;
[download]

push @{$fruits{$a}}, $b;
[download]

Update: Typo fixed. Thanks, Manav & Happy-the-monk.

push @{$fruits{$a}},$b ;
[download]



Manav

I see solutions to your problem, but since I don't see any explanation, I'll provide one.

Your problems lies with the fact that hashes can only have a single value at any given key. However, those value can be an array (or hash) reference which may contain a list of items.

One provided solution provided is:
push @{$fruits{$a}},$b
which is equivalent to:
$fruits{$a} = [] unless $fruits{$a}; # create array, save ref
push @{$fruits{$a}},$b

People suggested you could also form a string representing the list of values:
if (defined($fruits{$a})) {
   $fruits{$a} .= ','.$b;  # Add to existing list
} else {
   $fruits{$a} = $b;       # Start new list
}
This isn't good if you're planning on manipulating the individual values, but it good if you're just going to print out the list without modification.

One of the Perl docs refers to Array of Arrays (AoA). You should read up on those.

perldoc -f sort
[download]

Assuming you dont have undefined values or references as the values of @arr you can do this:

use Data::Dumper;
%fruits=();
@arr = ([1,1],[1,2],[3,5],[6,7]);
for (@arr)
{
    my ($k,$v)=@$_;
    if (ref $fruits{$k}) {
        push @{$fruits{$k}},$v;
    } else{
        $v=[$fruits{$k},$v] if (defined $fruits{$k});
        $fruits{$k}=$v;
    }
}

print Dumper(\%fruits);
__END__
$VAR1 = {
          '6' => 7,
          '1' => [
                   1,
                   2
                 ],
          '3' => 5
        };
[download]

Which is IMO a pattern perl programmers should be familiar with as it comes up fairly often. When only a few values will be lists the overhead of forcing each one to be an array can be unacceptable, but a hybrid works reasonably well. Imo generating stringified lists is not too desirable, but thats just me being a generalist, I suspect it could be a reasonable approach for some problems.

use strict;
use Data::Dumper;

my %fruits=();
my @arr = ([1,1],[1,2],[3,5],[6,7]);

for (0..$#arr) {
  $fruits{$_} = $arr[$_]
}

print Dumper(\%fruits);

__END__

STDOUT:

$VAR1 = {
          '1' => [
                   1,
                   2
                 ],
          '3' => [
                   6,
                   7
                 ],
          '0' => [
                   1,
                   1
                 ],
          '2' => [
                   3,
                   5
                 ]
        };
[download]

...
for (0..$#arr) { $fruits{$arr[$_]->[0]} = $arr[$_]->[1] }
...
# thus:

$VAR1 = {
          '6' => 7,
          '1' => 2,
          '3' => 5
        };
[download]

...
for (0..$#arr) { $fruits{$arr[$_]->[0]} = [@{$arr[$_]}[1..@{$arr[$_]}-
+1]] }
...
[download]