TMTOWTDI. The core module List::Util can be useful here:

#! perl
use strict;
use warnings;
use List::Util qw( first min );

my %h =
(
    17 => 2,
    23 => 5,
    42 => 1,
     7 => 3,
    35 => 4,
);

my $min_key = min keys %h;

printf "The smallest key   is %d, and its associated value is %2d\n",
        $min_key, $h{$min_key};

my $min_val = min values %h;

printf "The smallest value is %d, and its associated key   is %2d\n",
        $min_val, first { $h{$_} == $min_val } keys %h;
[download]

Output:

22:21 >perl 876_SoPW.pl
The smallest key   is 7, and its associated value is  3
The smallest value is 1, and its associated key   is 42

22:21 >
[download]

Hope that helps,

@sorted_positions = sort { $a <=> $b };
[download]

You can even do away with the intermediate array variable:

my $min = ( sort { $a <=> $b } keys %hash )[ 0 ];

If your hash is large, then using the sort function just to retrieve the smallest value is inefficient overkill. Use a variable, say $min, and set it to any key that you know to exist (or to a very large value such as 1e10 that you know will larger than at least one of the keys). Then you can simply do something like this:

my $min = 1e10;
for my $key (keys %h) {
     $min = $key if $key < $min; 
}
# now, $min is the smallest key. The associated value is simply $h{$mi
+n}
[download]

Now, depending on how you get the data in the first place, you could also maintain the $min variable when populating the hash, thereby avoiding to read all the keys once more.

But if you require 'special' hash, which naturally stores just one key/value pair, i.e. if new key/value pair is written into hash, one of them with greater key is deleted, you can implement your own hash using tie (http://perldoc.perl.org/perltie.html).