in reply to Re^2: non-scalar hash key
in thread non-scalar hash key

exactly

Replies are listed 'Best First'.
Re^4: non-scalar hash key
by McDarren (Abbot) on Jun 17, 2009 at 10:09 UTC
    ah, okay.

    Well, a reference to an array is a scalar. So, perhaps something like:

    my @foo = qw/1 2 3 4 5 6/; my %bar = ( \@foo => 'elephants', ); my $zonk = \@foo; print "$bar{$zonk}\n";
    prints:
    elephants

    Cheers,
    Darren :)

      The problem with your suggestion is that two arrays with the same elements will give two entries in the hash (because they will have different references). Consider:

      use strict; use warnings; use Data::Dumper; my @foo = qw/1 2 3 4 5 6/; my @fu = qw/1 2 3 4 5 6/; my %bar = ( \@foo => 'elephants', \@fu => 'zebras' ); print Dumper \%bar;

      Outputs something like:

      $VAR1 = { 'ARRAY(0x18ae770)' => 'zebras', 'ARRAY(0x17f0ec8)' => 'elephants' };

      And:

      my @foo = qw/1 2 3 4 5 6/; my @fu = qw/1 2 3 4 5 6/; my %bar = ( join ('',@foo) => 'elephants', join ('',@fu) => 'zebras' ); print Dumper \%bar;

      Outputs:

      $VAR1 = { '123456' => 'zebras' };

      citromatik

      A reference to another array with the same sequence of integers should point to the same hash entry.
Re^4: non-scalar hash key
by kdejonghe (Novice) on Jun 17, 2009 at 10:09 UTC
    I'm wondering why this is so difficult. If you take a string as the key to a hash, that string is actually an array of bytes, which is probably searched for by the C function strcmp. Now all I want is an array of which the elements are a tad larger (4 bytes iso 1 byte, because I need unsigned longs), and possibly another hash function (hashword in http://burtleburtle.net/bob/c/lookup3.c) and another comparison function similar to strcmp but working on integers.