sub next_key {
   my ($h, $key) = @_;

   my $best;
   foreach my $hkey (keys %$h) {
      if ( $hkey >= $key ) {
         if ( !defined($best) || $hkey < $best ) {
            $best = $hkey;
         }
      }
   }

   return $best;
}
[download]

If you use one of the many ordered hash implementations out there (Tie::IxHash, Tie::Hash::Indexed, Tie::Hash::Sorted, etc.) your code gets even simpler.

Untested example:

 use Tie::Hash::Sorted;

my %hash = (
    5189 => 63, 
    3213 => 9, 
    2357 => 3,
);

tie my %sorted_hash, 'Tie::Hash::Sorted',
    Hash         => \%ages,
    Sort_Routine => sub { $a <=> $b },
);


sub next_key {
    my ($h,$key) = @_;

    foreach my $hkey (keys %$h) {
        return $h->{$hkey} if $hkey >= $key;
    }
    return;
}

print next_key(\%hash,3000) . "\n";
[download]

You should be able to make a tied array.

A rough outline:

• Store your data in a hash, with keys as above.
• Keep an array of sorted keys.
• Your FETCH method will do lookups using a method like the above or even a binary search of your array of keys.
• Your STORE method will add an item to the hash, and resort the array.
• POP and SHIFT should remove and return the highest/lowest elements in the array.
• Not sure about UNSHIFT or PUSH. Should clearly define all the normal hash behavior.

Anyhow, this seems like a job for a tied array to me. Yeah you've got to do a lot of crud in the background to make it work, but once you are done, it becomes so easy to use.

 
tie my @numbers, 'Tie::Array::NextIndex',
    Members => {
        5189 => 63, 
        2357 => 3,
    };

my $value = $numbers[3000];  # value = 63
$numbers[3213] = 9;
$value = $numbers[3000];  # value = 9
[download]