In the following perl snippet:
my $a1 = [ qw(rock pop musical) ];
my $b1 = [ qw( mystery action drama )];
my $c1 = [ qw( biography novel periodical)];
my @a2d = (
$a1,
$b1,
$c1
);
The @a2d is an array which contain references to arrays.
My question is why the following print the same thing (musical)?:
print ${$a2d[0]}[2],"\n";
print $a2d[0][2],"\n";
I expected the second to print ARRAY or give an error since the elements of the array are refences
The $a2d[0] is an array reference. We can take this array reference and print out the 3rd entry:
my $ref = $a2d[0];
say ${ $ref }[2];
say $ref->[2];
These forms are equivalent. Now, we can do away with that intermediate variable, and get:
say ${ $a2d[0] }[2];
say $a2d[0]->[2];
If the dereference operator -> occurs between two subscripts, then it may be omitted as a shortcut:
say $a2d[0][2];
The arrow may be omitted when the left subscript is [...] or {...} and the right subscript it [...], {...} or (...).
This is also explained in perlreftut, which goes through these considerations in more depth. Reading that document should clear up many questions.
The dereference is implied when you tack on indexes.
$a2d[0][2]
is short for
${ $a2d[0] }[2]
aka
$a2d[0]->[2]
Rather than giving a syntax error, Perl provides a useful shortcut for a common operation.
Other examples: $aoa[$i][$j], $aoh[$i]{$k}, $hoa{$k}[$i] and $hoh{$k1}{$k2}.
