Re: sorting an array of hashes by the value of the keys

This will work for hashes with single keys

use Data::Dumper;

my @aoh = ({doc1=>2345}, {doc2=>1234}, {doc3=>5678}, {doc4=>4567});
my @sorted_aoh = sort { (%$a)[1] <=> (%$b)[1] } @aoh;

print Dumper \@sorted_aoh;

__output__

$VAR1 = [
          {
            'doc2' => 1234
          },
          {
            'doc1' => 2345
          },
          {
            'doc4' => 4567
          },
          {
            'doc3' => 5678
          }
        ];
[download]

HTH

_________ broquaint

Comment on Re: sorting an array of hashes by the value of the keys Download Code

Replies are listed 'Best First'.
Re: Re: sorting an array of hashes by the value of the keys by diotalevi (Canon) on Jun 19, 2003 at 15:34 UTC
I'll just expand on this so that it sorts on the key that starts with "doc" and ends with a number. This removes the "there must be only one key" requirement. `my @sorted_aoh = sort { (map /^doc(\d+)/ ? 0+$1 : (), keys(%$a))[0] <=> (map /^doc(\d+)/ ? 0+$1 : (), keys(%$b))[0] } @aoh;` [download] That happens to be suboptimal because every comparison is doing a potentially expensive operation to each hash. An improvement would be to use a Guttman-Rossler transform and then you get something really slick. `my @sorted_aoh = map $aoh[ substr $_, 5 ], sort map { # Retrieve the document number my $doc_key = (map /^doc(\d+)$/ ? 0 + $1 : (), keys %{$aoh[$_]})[0 +]; # Return the document number, a null and the original array index pack( 'N', $doc_key ) . "\0" . $_; } 0 .. $#aoh` [download]	[reply] [d/l] [select]

Replies are listed 'Best First'.

Re: Re: sorting an array of hashes by the value of the keys
by diotalevi (Canon) on Jun 19, 2003 at 15:34 UTC

I'll just expand on this so that it sorts on the key that starts with "doc" and ends with a number. This removes the "there must be only one key" requirement.

my @sorted_aoh = sort {
    (map /^doc(\d+)/ ? 0+$1 : (), keys(%$a))[0]
    <=>
    (map /^doc(\d+)/ ? 0+$1 : (), keys(%$b))[0]
} @aoh;
[download]

That happens to be suboptimal because every comparison is doing a potentially expensive operation to each hash. An improvement would be to use a Guttman-Rossler transform and then you get something really slick.

my @sorted_aoh =
map $aoh[ substr $_, 5 ],
sort
map {
    # Retrieve the document number
    my $doc_key = (map /^doc(\d+)$/ ? 0 + $1 : (), keys %{$aoh[$_]})[0
+];

    # Return the document number, a null and the original array index
    pack( 'N', $doc_key ) . "\0" . $_;
} 0 .. $#aoh
[download]

[reply]
[d/l]
[select]