If the file in question is not very big (e.g. less than, say, 50 MB on typical desktop machines of the last few years), and you have some reason to prefer your current ordering of the loops, it would be better to read the whole file content into a memory-resident array first, then use a nested "for" loop over that array, rather than repeating your "while" loop over the file contents.

The point is that file i/o is quite slow relative to memory-internal operations, so reading data from a file just once (rather than five times) just makes more sense.

This is bread and butter stuff for Perl and there are many better ways to achieve what you are after. Some are likely to be outside your current comfort zone so sift through the following and see what you like the look of:

use warnings;
use strict;

while (<DATA>) {
    chomp;
    my @numbers = unpack ("(a2)*", $_);

    print "unpacked: @numbers, ";

    @numbers = /(..)/g;
    print "matched: @numbers, ";

    @numbers = ();
    push @numbers, substr $_, 0, 2, '' while length $_;
    print "substr'd: @numbers\n";
}

__DATA__
0912203749
0725284648
0608294149
0622424347
1219303436
1729313449
1015162331
[download]

Prints:

unpacked: 09 12 20 37 49, matched: 09 12 20 37 49, substr'd: 09 12 20 
+37 49
unpacked: 07 25 28 46 48, matched: 07 25 28 46 48, substr'd: 07 25 28 
+46 48
unpacked: 06 08 29 41 49, matched: 06 08 29 41 49, substr'd: 06 08 29 
+41 49
unpacked: 06 22 42 43 47, matched: 06 22 42 43 47, substr'd: 06 22 42 
+43 47
unpacked: 12 19 30 34 36, matched: 12 19 30 34 36, substr'd: 12 19 30 
+34 36
unpacked: 17 29 31 34 49, matched: 17 29 31 34 49, substr'd: 17 29 31 
+34 49
unpacked: 10 15 16 23 31, matched: 10 15 16 23 31, substr'd: 10 15 16 
+23 31
[download]

I got sidetracked, but finally got back to this and took a look at your ideas, Grandfather. I'm a little disconcerted that your examples are simply reading each row into an array, in different ways. Thanks for the stylistic tips, but I'm not sure how that's relevant to what I'm trying to do (aggregate individual columns for analysis), but I guess I explained it badly. No matter. My stupid error was caught by Count Zero, and my awkward approach works fine for now. I may re-implement it using an array of references (my original idea), but I was aiming for a quick-and-dirty solution (not to worry: it's neither for work or a homework assignment!) :) . Thanks again.

Sans any context whatsoever, solving the immediate problem of reading and slicing up the file data seemed most helpful.

With the small amount of additional context you have added a fuller implementation would be:

use warnings;
use strict;

my @lines;

while (<DATA>) {
    chomp;
    push @lines, [unpack ("(a2)*", $_)];
}

__DATA__
0912203749
0725284648
0608294149
0622424347
1219303436
1729313449
1015162331
[download]

which pushes arrays of numbers into an array of lines creating a two dimensional array that you can use in subsequent processing:

for my $column (0 .. 4) {
    my @byColumn = map {$_->[$column]} @lines;
    print "@byColumn\n";
}
[download]

prints:

09 07 06 06 12 17 10
12 25 08 22 19 29 15
20 28 29 42 30 31 16
37 46 41 43 34 34 23
49 48 49 47 36 49 31
[download]

---

Perl is environmentally friendly - it saves trees