You're on the right track. What's happening is that when you open something the way you have it uses 'EOF' as the record. 'EOF' is usually after the newline character so it will leave you with one '\n' each time through the loop, thus your pattern trying to match '\n\n' fails. You really don't need a regex at all, all you need to do is chomp. If you only have one newline on a particular line the line will be deleted because the '\n' character is all that's there. This should work:

while(defined($rivi = <ROWS>)) {
     chomp $rivi;
     ....
}
[download]

Now chomp is removing all the '\n' characters so if you want each name on a separate line you'll need to add it back when you write to the file:

open(E,">>style1.txt") or die $!;
print E "$rivi\n"; # not understanding why you have $text here
close(E);
[download]

That will append the $rivi and a newline. Unless you have some other info in $text that you defined before this block of code I don't think you really need it (if you want the output you showed above)

I can think of two other ways to do this, if you're interested in those please reply

#!usr/bin/perl

open ROWS, "< style.txt" or die "Can't read from style.txt";

undef $/;    # no input record separator
$_ = <ROWS>; # copy the whole file to default var
close ROWS;

s/\s*(\w+)[\n\s]+([-.\d]+)[\n\s]+/$1 $2\n/g;
# substitute your described input format
# with requested output format in the whole text (now $_)

open E, "> style1.txt" or die "Can't write in style1.txt";
print E;
close E;
[download]

Here is a way that I used to solve your problem (be careful about the number of newlines and alike. It is a very fragile approach)

#!/usr/bin/perl -w
use strict;
my $a = join '', <DATA>;
$a =~ s/\n{2}(\d.\d+)/ $1/g;
$a =~ s/\n{2}/\n/g;
print $a;

__DATA__

Alan

2.13

Mary

2.16
[download]

This prints:

Alan 2.13
Mary 2.16
[download]

I hope this is of any use.

One of these there-are-many-ways-to-solve-this perl questions.

{
local($/)=undef;
map { s/([A-Za-z]+)\W+([\d\.]+)/print STDOUT "$1 $2\n"/ge } <DATA>;
}
exit 0;
__DATA__
Alan
 
2.13
 
Mary
 
2.16
[download]

   perl -pi -we 'chomp; s/^([A-Za-z])/$1 /; s/^([\d.]+)/$1\n/' style.t
+xt
[download]

Finally the "chomp" solution was easy to put in my code as I already had it and removed all unnesseray stuff. But what about if the next line contains data I wish also to have left also?

Like original file

Mary

2.12 Some text here we want too

David

3.13 Even more text here
[download]

Mary 2.12 Some text here we want too

David 3.13 Even more text here
[download]

Effectively what you want to do is remove any blank lines and then join every second remaining to it's predecessor. So... (untested)

#open INPUT and OUTPUT
while (<INPUT>) {
    chomp;
    next if m/^\s+$/;
    my $nextline;
    chomp($nextline = <INPUT> ) while $nextline !~ m/^\s+/;
    print OUTPUT $_,' ',$nextline;
}
#close INPUT and OUTPUT
[download]

---

What's this about a "crooked mitre"? I'm good at woodwork!

Replacing the regex in my first answer with that one should work:

s/\s*(\w+)[\n\s]+([^\n]+)[\n\s]+/$1 $2\n/g;
[download]