You never get a copy of the array referenced by $data2, not even the first time through the loop.

 
my @array;
foreach (0..2) {
        my $d = $data1->[$_];
        $d->{sizes} = $data2;
        push(@array,$d);
}
print Dumper(\@array);

$VAR1 = [
          {
            'food' => 'apple',
            'shape' => 'round',
            'sizes' => [
                         {
                           'small' => 'bunny',
                           'big' => 'cow'
                         },
                         {
                           'small' => 'mouse',
                           'big' => 'horse'
                         }
                       ]
          },
          {
            'food' => 'pear',
            'shape' => 'square',
            'sizes' => $VAR1->[0]{'sizes'}
          },
          {
            'food' => 'grape',
            'shape' => 'oval',
            'sizes' => $VAR1->[0]{'sizes'}
          }
        ];
[download]

        push(@array,clone($d));
[download]

So, why does it "work" on the first pass (without using clone) and then from there I just get a shallow copy?

Short answer: it did *not* "work" on pass#1 any differently than pass#2 or pass#3. That fact is might not be evident from Data::Dumper output until one obtains mastery of Perl references (via perlreftut and perlref).

Detailed explanation:
If you say

my @foo = ( 1, 2, 3 );
my $arrayref = \@foo;
[download]

, and you wanted to explain to a fellow programmer what $arrayref pointed to, you would (of course) simply say "the array named @foo".


If instead you had coded

my $arrayref = [ 1, 2, 3 ];
[download]

, then $arrayref would still point to equivalent data, but you could *not* describe it as a reference to a particular array named @something. The array is nameless, floating in memory as long as anyone has a reference to it, and vanishing once the last reference goes away. It is anonymous.

Furthermore, if I say

my @foo = ( 1, 2, 3 );
[download]

, then I would be (close enough to) accurate to say that the three numbers "live" in @foo, and that the 2 is stored between the 1 and the 3. If instead I code

my @foo = ( [1], [2], [3] );
[download]

, I would be far less accurate to say that the three anonymous arrays "live" in @foo, or that the anonymous array containing 2 is stored between the anonymous array containing 1 and the anonymous array containing 3.

That distinction is important. Arrays cannot hold arrays or hashes; Arrays can only contain scalars. Hash values can only contain scalars. The three anonymous arrays "live" out in space, only the *references* to them get stored (in an ordered fashion) in the @foo array.

When Data::Dumper says "$VAR1->[0]{'sizes'}" in your second two passes, it does so because that string of code is the closest thing to a name it has to describe the anonymous hash it already described in the first pass. Data::Dumper cannot print simple repetitions of the anonymous hash from the first pass, because that would be describing 3 separate anonymous hashes, each (coincidentally) containing the same data elements, yet each floating in its own separate place in memory.

Perhaps this simpler example will help.

my $arrayref1 = [ 9, 10, 'a big fat hen' ];

my $arrayref2 = [
    $arrayref1,
    $arrayref1,
    $arrayref1,
];

print "Only showing super-structure:\n";
print Dumper $arrayref2;
print "Showing both sub-structure and super-structure:\n";
print Dumper $arrayref1, $arrayref2;
[download]

produces this output:

    Only showing super-structure:
    $VAR1 = [
              [
                9,
                10,
                'a big fat hen'
              ],
              $VAR1->[0],
              $VAR1->[0]
            ];
    Showing both sub-structure and super-structure:
    $VAR1 = [
              9,
              10,
              'a big fat hen'
            ];
    $VAR2 = [
              $VAR1,
              $VAR1,
              $VAR1
            ];
[download]

When Data::Dumper says "$VAR1->[0]{'sizes'}" in your second two passes, it does so because that string of code is the closest thing to a name it has to describe the anonymous hash it already described in the first pass.

Ok, I am getting my head around this now... thanks for taking the time to explain this in such detail. Much appreciated... Cheers!