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

my $TEXT= do { local $/; <DATA> };
my $result;
if(  $TEXT =~ /(.*?abc)/g  ) {
    $result= $1;
    while(  $TEXT =~ /\G(\D+)([\d.\-]+)/g  ) {
        $result .= $1 . "GOVALUE";
    }
    $result .= substr( $TEXT, pos($TEXT) );
}
print $result;
__END__
                    abc(" 0.123511,  0.074997,  0.103143,  0.153955,  
+0.260586",\
                            " 0.068394,  0.082183,  0.110329,  0.16114
+2,  0.267773",\
                            " 0.081887,  0.095677,  0.123822,  0.17463
+5,  0.281266",\
                            " 0.111534,  0.125324,  0.153470,  0.20428
+2,  0.310914",\
                            " 0.165731,  0.179521,  0.207666,  0.25847
+9,  0.365110");
[download]

Makes one wish you could do incremental s///g (like m//g in a scalar context), though we'd need something like an extra option to request such. s///gc or s///G (belg4mit's) anyone? Of course, it would probably be less efficient than the above code (requiring the entire end of the string to be copied/moved multiple times), but it would be less error-prone.

You know, you can use \G in s///g after doing a //g.

$TEXT =~ /abc/g;
$TEXT =~ s/\G(\D+)([\d\.\-]*)/$1GOVALUE/g;
[download]

---

Caution: Contents may have been coded under pressure.

Did you test this? I'm not claiming you are wrong, but my quick test of it showed it not changing anything:

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

my $TEXT= do { local $/; <DATA> };
if( $TEXT =~ /abc/g ) {
    $TEXT =~ s/\G(\D+)([\d\.\-]*)­/$1GOVALUE/g;
}
print $TEXT;
__END__
                    abc(" 0.123511,  0.074997,  0.103143,  0.153955,  
+0.260586",\
                            " 0.068394,  0.082183,  0.110329,  0.16114
+2,  0.267773",\
                            " 0.081887,  0.095677,  0.123822,  0.17463
+5,  0.281266",\
                            " 0.111534,  0.125324,  0.153470,  0.20428
+2,  0.310914",\
                            " 0.165731,  0.179521,  0.207666,  0.25847
+9,  0.365110");
[download]

The output was the same as the input.

- tye        

    use Regexp::Common;
    $text =~ s/$RE{num}{decimal}/GOVALUES/g;
[download]

Do you want the entire set of numeric values replaced with a single character string, "GOVALUE?"

And when you say "repeat my code" do you mean replicating it, in line, with essentially identical code?

In other words, please help me understand (I see there are others who may have provided the help you wanted) by an example or more precise description of the desired output (and the method of your "repeat my code."

s/     start "substitution" -- not exactly "find and replace"
    abc    # match (but don't capture or keep) "abc"
    (\D+)    # capture one_or_more non-digits followed by a space
    ([\d\.\-]*)    # capture (as $2) zero_or_more digits, dots, slashes** or minus-signs (fol by a space)
    (.*)/    # On the data provided, capture (as $3) anything and everything except a newline, starting at the first comma after the first decimal value

** or are you trying to escape the hyphen in class? if so, that's not necessary as the first or last element because if its the last thing in a class it can't be a range operator

and replace the original $text with:
"values" followed by the non-digits in $1 "(open_paren, dquote)" followed by "GOVALUE" followed by the contents of $3 -- everything not previously captured in $1 or $2.

values(" GOVALUE,  0.074997,  0.103143,  0.153955,  0.260586",
                 " 0.068394,  0.082183,  0.110329,  0.161142,  0.26777
+3",
                 " 0.081887,  0.095677,  0.123822,  0.174635,  0.28126
+6",
                 " 0.111534,  0.125324,  0.153470,  0.204282,  0.31091
+4",
                 " 0.165731,  0.179521,  0.207666,  0.258479,  0.36511
+0")
[download]

...if your data is qq'ed into the value assigned to $text. ie,

#!c:/perl/bin -w

$text = qq ( abc(" 0.123511,  0.074997,  0.103143,  0.153955,  0.26058
+6",\
                  " 0.068394,  0.082183,  0.110329,  0.161142,  0.2677
+73",\
                  " 0.081887,  0.095677,  0.123822,  0.174635,  0.2812
+66",\
                  " 0.111534,  0.125324,  0.153470,  0.204282,  0.3109
+14",\
                  " 0.165731,  0.179521,  0.207666,  0.258479,  0.3651
+10") );

$text =~ s/abc(\D+)([\d\.\-]*)(.*)/values$1GOVALUE$3/g;

print "$text\n";
[download]

However, while trying to make certain I wasn't mistating (was I?) re escaping a initial or final hyphen in a class, I started playing with variant forms, such as:

$text2 = 'foo|abc-./|foo ';

# $text2 =~ s%(foo|\w+)([//.-]+)(.*)%$1 :i: $2 :i: $3%g;

# output is:  ++foo|abc :i: -./ :i: |foo ++
# captures the hyphen, dot and slash in $2 with ONLY the fwd_slash esc
+aped 
# and capturs "|foo " in $3  

# $text2 =~ s%([\w|]+)([.-]+)(.*)%__$1_:_$2_:_$3__ : second variation%
+g;

# output is:  ++__foo|abc_:_-._:_/|foo __ : second variation++
# hyphen, dot still captured in $2 with no escaping at all inside the 
+class

$text2 =~ s%([\w|]+)([.-]+/)(.*)%$1:$2:$3 : third variation%g;

# output is: ++foo|abc:-./:|foo  : third variation++
# Moved the fw_slash (unescaped, using % as delimiters) into the secon
+d capture 
# expression... and fw_slash turns up correctly in $2

print "\t++$text2++\n";
[download]

But the alternates above go all around the barn. Any hints, please, on a better way to do a demo on this (narrow) matter?

Never wonder that out loud

s/(\D+)([\d\.\-]*)/values$1GOVALUE$2/g;
[download]