Re^2: mem usage

[halfcountplus]

Yeah, that's Fisher-Yates (aka Knuth #12 or something). It works in place and I've never had any problems with it,* dunno why you think it's "not very good". Kind of a wacky place to put i-- but that's pretty trivial, style wise.

*(eg, just shuffled those 12 millon lines, no errors, since I am feeding that into something else I will notice if something gets duplicated et. al.)

[BrowserUk]

You can trim your shuffle a little by omitting the line:         next if $i==$j;.

Swapping an item with itself doesn't affect the algorithm's fairness, and doing it once costs less, than testing n times and avoiding it once.

And you save a little more by avoiding the list assignment:

my $tmp = $array[ $i ]; 
$array[ $i ] = $array[ $j ]; 
$array[ $j ] = $tmp;

Doesn't look as nice though.

[ikegami]

Including cleaning up the counter,

sub shuffle {
    my $array = shift;
    my $i = @$array;
    while ($i) {
        my $j = int rand($i--);
        my $t = $array->[$i];
        $array->[$i] = $array->[$j];
        $array->[$j] = $t;
    }
}

[halfcountplus]

You can trim your shuffle a little by omitting the line: next if $i==$j;

Good point!