pzbagel was right in saying that the ordering of hashes in Perl is not deterministic: nothing guarantees that you always get stuff out in the same order as you put them in (in fact, given the nature of the hashing algorithm, you are virtually guaranteed that you get them out in a different order than they went in). If you want to guarantee a consistent order (though not necessarily the initial order), use sort on the list returned by keys %hash.

However, this seems to be more of a case for arrays than hashes. As you say, all your hashes have identical keys, both in number and name. If you want to associate some kind of descriptive name to the array elements, keep a seperate array of names. Something like this:

my @names=qw/value1 value2 value3/;
my @array1=(1, 2, 3);
my @array2=(4, 5, 6);
[download]

my @AoA;
push @AoA, [ 1,2,3 ];
push @AoA, [ 4,5,6 ];
#Then use them as such:
$AoA[0][0]; # 1st element of first array
$AoA[1][0]; # 1st element of second array
[download]

CU
Robartes-

As is noted in the perl documentation regarding hashes, the order that hash values will come out of a hash is not guaranteed. You should use keys() and then iterate over the keys (getting the value from each hash) to do your dotproducts. This will guarantee you are operating on the correct pairs of values.

Cheers

I agree with robartes, using hashes for this application is basically the wrong way to do it. It would seem that the only reason you are using the hash is to give meaningful names to the values. An alternative is to use constant names for the array elements via the constant pragma.

However, its not clear from your snippet whether you are generating the hashes (in which case you could re-work the hashes to arrays) or whether these are being generated by a module which would make life harder.

You might also look at Tie::IxHash which will always return the values in the order the keys were created, but there is a performance penalty associated with this which, if the hashes are small might outweight the benefits of avoiding sorting. And again, if the hashes are being generated in a module, it could be awkward to persuade it/them to use Tie::IxHash.

my @keys = keys %{$database[0]};  #optionally use sort keys
my $pdl1 = pdl (@{$database[0]}{@keys});
my $n1 = norm $pdl1;
my ($pdl2, $n2, $dotproduct, $d, @angles);
for (0..$3database) {
      $pdl2 = pdl ((@{$database[$_]}{@keys});
      $n2 = norm $pdl2;
      $dotproduct = inner ($n1, $n2);
      $d = dotproduct->sclr();
      $angles[$_] = (180/3.1415926)*acos($d);
}
[download]