perl -pni.bak -e "s/1 (\w{1,2}v\w{1,2})/2 $1/" filename
[download]

c:\> for %f in (*.html) do perl -pni.bak -e "s/1 (\w{1,2}v\w{1,2})/2 $
+1/" %f
[download]

perl -pi.bak -e 's/\b(\w{1,2}v\w{1,2})\b/g<a href="$1.html">$1<\/a>/" 
+filename
[download]

Hmm, I'm not sure if that's exactly what he wants.. I think he is looking to add it multiple times (which would ofcourse only mean that the replacement part of the s/// changes into $1$1$1... (or $1 x 3) (for example).

simple, you should be able to use *.html as filename...

Small note, why do you use '-pn'? -p: print, -n: don't print... -p should do just fine

hmm the more ive read on this the more im unsure of what the solution would look like. somewhere im expecting to see something that for each
$1= *v* <v> gives code1 $1 code2 $1 and then so on for each *v* in each file... - (steamerboy)

well the 'names' i mentioned arent links - theyre names for values that get recorded. The code for the checkboxes is like
(code value=1 name =xvx) (code value=2 name= xvx)... and so on for each checkbox. So the idea is for the code to get any ? v ? and replace it with itself embedded in the code - saving me 500 years of time! Thanks for the help - much aprreciated! (steamerboy)

It would be much easier if you gave real examples of the input and the expected output...

#use strict;
#use warnings;

my ($dirname, $file, %words);

opendir(DIR, ".") or die "can't opendir $dirname: $!";

while (defined($file = readdir(DIR))) {
    if ($file =~ /\.html?/) {
        open my $F, $file or die "Could not open $file\n";
        while (<$F>) {
        #do stuff
        s/
            (?<=                  # Look back-group
             weak                # Match the word weak
             <INPUT TYPE=radio NAME=     # tag
             (                   # Capuring group: $1 / \1
               \w{2}\d\w{2}      # 
             )                   # End-capture
             VALUE=\d>           # tag
         
             (?:                 # Non capturing group
               <INPUT TYPE=radio NAME=     # tag
               \1                # It's possible that this should be $
+1, as I said, the code is untested.
               VALUE=\d>
             ){6}                # There are 7 input tags, one already
+ matches, so 6 to go.
             )                   # End look-back
             (?=                 # Look-ahead-group
               strong
             )                   # End look-ahead
           /
             the text you want to insert
           /ixg;
        }
        close $F;
        }
}
closedir(DIR);
[download]

Note that your regex also matches

• .htmanagerrc
• foo.htm.old
• bar.htmx
Better written as:

if ($file =~ /\.htm(l)?$/) {
[download]



holli, /regexed monk/

(Update) code at the top of this post does NOT work

perl -pi.bak -e 's/(?<=weak<INPUT TYPE=radio NAME=(\w{2}\d\w{2}) VALUE
+=\d>(?:<INPUT TYPE=radio NAME=\1VALUE=\d>){6})(?=strong)/insert($1)/e
+ig;
BEGIN { sub insert { my $name = shift; return "value 8, value 9, value
+ 10"; } }' *.html
[download]

Explenation of regex:

  s/
    (?<=                  # Look back-group
     weak                # Match the word weak
     <INPUT TYPE=radio NAME=     # tag
     (                   # Capuring group: $1 / \1
       \w{2}\d\w{2}      # 
     )                   # End-capture
     VALUE=\d>           # tag
 
     (?:                 # Non capturing group
       <INPUT TYPE=radio NAME=     # tag
       \1                # It's possible that this should be $1, as I 
+said, the code is untested.
       VALUE=\d>
     ){6}                # There are 7 input tags, one already matches
+, so 6 to go.
     )                   # End look-back
     (?=                 # Look-ahead-group
       strong
     )                   # End look-ahead
   /
     the text you want to insert
   /ixg;
[download]

Big note, the code in the explenation will not work properly, because the x-modifier is in use and I did not escape the whitespace!

Update: after reading the reply of Anonymous Monk, I decided to test it... the variable lookbehind error comes from \1. This can be fixed by using (?=\1).{5} instead of \1, but then another errors shows up: 'Lookbehind longer than 255 not implemented'. This makes it impossible to use a look-behind...

So there is only one thing left to do and that is making it a 'real' group of it...

Then the short (and tested (or atleast on simple data)) version would become:
perl -pi.bak -e 's/(weak<INPUT TYPE=radio NAME=(\w{5}) VALUE=\d>(?:<INPUT TYPE=radio NAME=\2 VALUE=\d>){6})(?=strong)/$1 . insert($2)/eig; BEGIN { sub insert { my $name = shift; return "value 8, value 9, value 10"; } }' *.html

The differences:

• A capturing group is used,
• $1 and \1 changed into $2 and \2,
• $1 . insert($2),
• The input pattern is corrected (I used \d where it should have been 'v')
• A space is added between \2 and value (it was missing)

hmm i get this error; Variable length lookbehind not implemented in regex; marked by <-- HERE in m/(?<=weak<INPUT TYPE=radio NAME=(\w{2}\d\w{2}) VALUE +=\d>(?:<INPUT TYPE=radio NAME=\1VALUE=\d>){6})(?=strong) <-- HERE / at -e line 2. what does this mean? Thanks for the help! Great stuff. (steamerboy)