http://qs321.pair.com?node_id=618804


in reply to removing the goto

I had the benefit of a direct discussion with the OP about this, which brought much clarity as to his intentions. He has an array of ~4000 elements and wants to select five non-duplicates out of there. So basically, his code can be reduced to this: 
use List::Util 'shuffle';
my %unique = map { $_ => 1 } @weighteddiv;
my @shuffled = shuffle(keys %unique);
@selected = @shuffled[0..4];

[download]
I noticed he was randomizing the array, then selecting random indices from there, which was not only reduntdant but created the issue of potentially selecting the same element.

Update: The big array is *not* already internally unique after all, so I had to add the line to unique it.

Update 2: My solution is no longer relevant, after the OP updated with more information. However, I'm leaving my code as-is because it might help a future monk looking for a similar solution.

---
It's all fine and dandy until someone has to look at the code.

Replies are listed 'Best First'.
Re^2: removing the goto
by jdporter (Chancellor) on Jun 01, 2007 at 18:27 UTC 
    use List::Util 'shuffle';
use List::MoreUtils 'uniq';

my @shuffled = shuffle( uniq @weighteddiv );

@selected = @shuffled[0..4];

    [download]
[reply]
[d/l]
      It might be more appropriate to do the shuffle before the uniq if the OP wants the duplicate values to have a greater chance of being selected - as suggested by the name @weighteddiv. 
      use List::Util 'shuffle';
use List::MoreUtils 'uniq';
use strict;
my @weighteddiv = (
          (1..50),
          (1..10) x 10,
      );

my @shuffle1 = shuffle( uniq @weighteddiv );
my @shuffle2 = uniq( shuffle @weighteddiv );

printf "shuffle1: %s\n", join(', ', sort {$a <=> $b} @shuffle1[0..10])
+;
printf "shuffle2: %s\n", join(', ', sort {$a <=> $b} @shuffle2[0..10])
+;

      [download]
      Output: 
      shuffle1: 2, 5, 16, 17, 23, 25, 26, 27, 39, 42, 43
shuffle2: 1, 3, 4, 5, 6, 8, 9, 10, 13, 40, 50

      [download]
      This approach is appropriate for the current implementation of uniq, which preserves the original order while removing duplicates.
[reply]
[d/l]
[select]

In Section Seekers of Perl Wisdom