It would look something like this:

my @a = ([1,2], [3,4]);
my @b = sort {

    $a->[0] <=> $b->[0] || # the result is -1,0,1 ...
    $a->[1] <=> $b->[1]    # so [1] when [0] is same

} @a;
[download]

Check out sort for the full story. And, you'd probably have to maintain some kind of global setting for the sub (indicating which columns to sort by) or call the sub from the sort block indicating it there. Although, calling functions from the sort block can get surprisingly slow. The worst case is O(n*log(n)) I believe and calling functions in perl is among the slowest thing it does.

I'm sure you'll figure it out.

UPDATE: NetWallah, sorry, no bugs here. This was not intended to be a complete solution. What's the fun in having someone else write it all out for you?

Worked like a CHARM! Thanks! That was also MUCH easier than I would have ever imagined. Best Regards



After all this is over, all that will really have mattered is how we treated each other.

use Test::More qw/no_plan/;

my $unsorted = [
    [ qw/1  3  5  6  2/ ],
    [ qw/3  4  5  6  7/ ],
    [ qw/1  2  3  4  5/ ],
    [ qw/5  6  7  8  8/ ],
    [ qw/1  2  3  4  5/ ],
    [ qw/2  2  2  2  2/ ],
    [ qw/1  1  2  4  5/ ],
    [ qw/2  3  4  5  6/ ],
];

my $expected = [
    [ qw/1  1  2  4  5/ ],
    [ qw/1  2  3  4  5/ ],
    [ qw/1  2  3  4  5/ ],
    [ qw/1  3  5  6  2/ ],
    [ qw/2  2  2  2  2/ ],
    [ qw/2  3  4  5  6/ ],
    [ qw/3  4  5  6  7/ ],
    [ qw/5  6  7  8  8/ ],
];

my $sorted = sortSub($unsorted, 0, 1, 3);

is_deeply $sorted, $expected;

sub sortSub {
    my @array = @{ +shift };

    my @sorted = sort {
        for my $ix ( @_ ) {
            my $cmp = $a->[$ix] <=> $b->[$ix];
            return $cmp if $cmp;
        }
        return 0;
    } @array;
    return \@sorted;
}
[download]

Update: Fixed a mistake spotted by dbmathis Narveson.

Nice, but shouldn't

for my $ix ( 0 .. @_ ) {
[download]

instead be

for my $ix ( @_ ) {
[download]

?

Yes, you're right. I'll update the node now.

Thanks

How would I make this work if I am passing in @unsorted rather than $unsorted?



After all this is over, all that will really have mattered is how we treated each other.

I had already indicated that in the comment of my previous node. Here it is again:

my @sorted =  sort NetWallahSort   @unsorted;
[download]

Note : I have updated "sub NetWallahSort" in the posted code, to include ikegami's suggestion, returning "0" instead of "1" when sort elements are equal. Please use a current copy of that code.

     "As you get older three things happen. The first is your memory goes, and I can't remember the other two... " - Sir Norman Wisdom

Hi Funk, I like your solution and I am able to get it working fine in test, however in my existing code I will be passing a list of array references within an array to the routine (@unsorted).

For example here's the part of my code where I am loading the array:

while ( @row = $sth->fetchrow_array ) {
    push @x, [ @row ];
}
[download]

    $rindex = 0; $i = 1;
    do {
        foreach ( @columns ) {
            if ( $_ <= $qstart_column - 2 ) {
                $unsorted[$rindex][$_] = $x[$i][$_];
            } else {
                $unsorted[$rindex][$_] = $x[$i][$_ - $nun_of_questions
+];
            }
        }
        $rindex += 1; $i += $nun_of_questions;
    } while ( $i <= $#x );
[download]

Right now when run your code in my code I get the following:

Use of reference "ARRAY(0x10b5810)" as array index at /home/dmathis/qw
+est_qqs/transpose-qqs-dev line 367.
Use of reference "ARRAY(0x10b5810)" as array index at /home/dmathis/qw
+est_qqs/transpose-qqs-dev line 367.
Use of uninitialized value in numeric comparison (<=>) at /home/dmathi
+s/qwest_qqs/transpose-qqs-dev line 367.
Use of uninitialized value in numeric comparison (<=>) at /home/dmathi
+s/qwest_qqs/transpose-qqs-dev line 367.
................
................
................
[download]

I just need to figure out what to alter in your code to handle my array.

Best Regards