my @data
while (my $line = <TEMPFILE>) {
        my @elements = $line;
        push @data, \@elements;    # need to store as ref
    }

@data = sort { $$a[2] cmp $$b[2]  or  $$a[1] <=> $$b[1] } @data or die
+ "can't sort";



foreach my $item (@data) {
    print TEMPFILE @$item;
    print TEMPFILE "\n";
}
[download]

Edit by tye

Sidenote: can you please please do us a favour and read Writeup Formatting Tips. Your posts are just about unreadable as is. Specifically, at least learn about the <code> tag.

Remove the or die "Can't sort". The or forces scalar context on its left operand, which is not what you want.

Makeshifts last the longest.

Hmm. I don't see any map in the snippet you posted. Actually, unless it got skipped in the copy paste, it doesn't seem like you're separating your input into columns at all.

Makeshifts last the longest.

You don't need the map function, rchou2. You took care of everything by using the for() loop. However, you did not split the data into fields.

Here's one way to do what you are asking, that borrows most of the syntax preferences from your example:

#!/usr/bin/perl -w
use strict;

rename 'tempfile.txt', 'tempfile.txt~'; # save a backup

open TEMPFILE, "<tempfile.txt~" or die $!;

my @data;
my @newcolumn = qw(d 
                   e
                   f);
my $incr = 0;
for my $line (<TEMPFILE>) {
        
        my @elements = split(/\s+/g, $line); # split the line on all w
+hitespace
        push @elements, $newcolumn[$incr++]; # add a new column
        push @data, \@elements;              # need to store as ref
}
close TEMPFILE;

@data = sort { $$a[2] cmp $$b[2]  ||  $$b[1] <=> $$a[1] } @data # or d
+ie "can't sort";

$,    = " " ;         # separate output fields by a space

open OUTFILE, ">tempfile.txt" or die $!;

foreach my $item (@data){
    print OUTFILE "@$item\n";
}
close OUTFILE;
__END__
[download]


mkmcconn