Re: hash sort

[johngg]

The one-to-one correspondence of the %hash1 and %hash2 keys makes me a little suspicious, where is the point of having both hashes if that's the whole of the data? I'm going to hazard a guess that %hash2 is larger than %hash1, the latter being used to select a sample from the former. I call them %data and %pick in the script below which uses rand to generate the data (and srand so that it is consistent between runs).

use strict;
use warnings;

srand 12345;

my %pick =
   map { $_ => 1 }
   1 .. 5, 8, 12 .. 16, 19, 22, 25, 37 .. 41, 62, 75, 77;

my @letters = ( q{a} .. q{e} );

my %data =
   map { $_ => { key => $letters[ rand @letters ] } } 1 .. 80;

printf qq{key %2d and %s\n}, unpack q{xNXXXXXa} for
   sort
   map { pack q{aN}, $data{ $_ }->{ key }, 0 + $_ }
   keys %pick;

The output.

key 13 and a
key 15 and a
key 16 and a
key 25 and a
key 41 and a
key  1 and b
key  3 and b
key 22 and b
key 40 and b
key 12 and c
key 38 and c
key 62 and c
key 75 and c
key  4 and d
key  5 and d
key 19 and d
key 37 and d
key  2 and e
key  8 and e
key 14 and e
key 39 and e
key 77 and e

I hope this guess bears some relation to the real problems and is helpful.

Update: A perhaps more readable alternative to the "skip one - read four - skip back five - read one" unpack template would be this:-

printf qq{key %2d and %s\n}, reverse unpack q{aN} for
   sort
   map { pack q{aN}, $data{ $_ }->{ key }, 0 + $_ }
   keys %pick;