This should work:

print join "\t", split ' *(?=[^0-9,])', '15,567 -25,324-45,234';;
15,567  -25,324 -45,234
[download]

As a one-liner:

perl -nle"print join qq[\t], split ' *(?=[^0-9,])'" < bad.file > good.
+file
[download]

Don't forget to swap 's and "s if you're on *nix.

I don't know how I managed to get it wrong but after an hour of trying I made them both work :) Thank you very much!

Just a guess, are the columns fixed width? If so, unpack can help.

while (<DATA>) {
    chomp;
    my @fields = unpack('a7 a7 a7');
    print join(" ", map {"\"$_\""} @fields), "\n";
}
__DATA__
 15,567-25,324-45,234 
-13,345 53,562 13,452 
 -7,521-22,454-54,671
[download]

Outputs:

" 15,567" "-25,324" "-45,234"
"-13,345" " 53,562" " 13,452"
" -7,521" "-22,454" "-54,671"
[download]

Hi and thanks for your answer, it's the one I think I understand the most, but it's very kind that you all took the time and tried to help. I'll also try to understand the other two answers hopefully.

I wasn't successful implementing this but I'm wondering if it has to do with the fact that I haven't really given you a clear picture of what my data looks like, so this file looks like this - this is probably a pretty silly way to show you but here's a link to a screenshot of my data https://www.dropbox.com/s/68xjbspn47jzmx9/Untitled.png?dl=0 , sorry I know this is really random but I have no clue how to properly format the numbers so that you can see.

As you can see I have 4 columns, the last one is always 1 digit and the rest always have 3 decimal digits. Can your solution (with some editing) still be implemented on my data?

Can't you just copy and paste some of your data here inside <code> tags? Those will preserve formatting.

Anyway, from this small sample it does look like your columns are fixed width. The first one looks to have either 7 or 8 characters, the second and third 8 characters - try an unpack pattern of "a8 a8 a8" and see what you get. Adjust and extend this pattern as is appropriate for your input data, for example to add another column "a8 a8 a8 a6" and so on.

For a tutorial see "Packing Text" in perlpacktut.

Try:

while(<$FILE>) {
  chomp ($line = $_);
#  @columns = $line =~ m/\s*(-?\d+)/g; 
  @columns = $line =~ m/\s*(-?[\d,]+)/g;
}
[download]

Update: thanks to Lotus1 for the correction. I had it right on my computer, but failed to update the post.

Hi GotToBTru, \d does not match commas so your program splits up the numbers. If you add the comma in a character class it works.

use warnings;
use strict;

while(<DATA>) {
  chomp (my $line = $_);
  my @columns = $line =~ m/\s*(-?\d+)/g;
  print "@columns\n";
}

__DATA__
15,567 -25,324-45,234
 15,567-25,324-45,234 
-13,345 53,562 13,452 
 -7,521-22,454-54,671
[download]

Output:
15 567 -25 324 -45 234
15 567 -25 324 -45 234
-13 345 53 562 13 452
-7 521 -22 454 -54 671
[download]

Here it is with the comma added.

use warnings;
use strict;

while(my $line = <DATA>) {
  my @columns = $line =~ m/\s*(-?[,\d]+)/g;
  print "@columns\n";
}

__DATA__
15,567 -25,324-45,234
 15,567-25,324-45,234 
-13,345 53,562 13,452 
 -7,521-22,454-54,671
[download]

Output:
15,567 -25,324 -45,234
15,567 -25,324 -45,234
-13,345 53,562 13,452
-7,521 -22,454 -54,671
[download]

Yes, exactly, this works correctly. Thank you both though, I really appreciate the help and explanations, this is an amazing website.

Hi angela2

I can see that you've already had some great responses - some more easily understood than others; so for posterity here are a couple of one-liners similar to what BrowserUk suggested here Re: problem with splitting file on whitespace: how to circumvent inconsistent formatting through file that also take into account tab delimited columns and numbers with decimal points.

one-liner to print all columns

one-liner to print column 1

...use $F[n-1] to access column n, e.g. print $F[1] will print the value for column 2.

quick explanation of command-line flags

• -a and -F splits line text into the @F array; -a tells it to do the splitting and -F is used to specify the delimiter to split on - the default delimiter is space
• -w sets the warning flag similar to the use warnings pragma
• -n along with -p are the more commonly used command line switches both concerned with reading <ARGV>. -n provides us with the implicit construct

  while (<>) { "...your code here ..." }

• -l sets the output record separator $\ to the input record separator $/ - which is \n by default. I often use it to preclude the need to add \n in my print statements
• -e the most commonly used switch - it tells perl not to load and run a program file but to run the text following the -e as a program

See perlrun for a more info as I have only scratched the surface...

Best Wishes,
shadowsong

This is good stuff, thanks!