If you need assistance, please show us the code you have already written and explain how it fails to do what you need.

Please use <code>...</code> tags to show us your code, data and the output your code produces.

I need help to replace comma in the matched string boundaries.

String like (12,4666,{xyx,yyy,ab1{1213,34343},3433,xy1[{3232,989},{8737,933}],3343,pq1[{1212,989}],98383},8383,xty)

I need to replace comma in sub structure ab1{1213,34343} to modify as ab1{1213|34343} and replace comma in sub List of Structure xy1[{3232,989},{8737,933}] as xy1[{3232|989}|{8737|933}]

and result as (12,4666,{xyx,yyy,ab1{1213|34343},3433,xy1[{3232|989}|{8737|933}],3343,pq1[{1212|989}],98383},8383,xty)

Otherway i need to replace comma in the substring starting with <string>{ and ending with } and <string>[{ and ending with }]

Please assist. Thanks in advance

(Better typography in the node above makes the problem more clear, but use of  <code> ... </code> tags would have been better yet. Please see Markup in the Monastery and Writeup Formatting Tips.) Here's one approach (needs Perl version 5.10+ for \K and //):

c:\@Work\Perl\monks>perl -wMstrict -le
"use 5.010;
 ;;
 my $s = '(12,4666,{xyx,yyy,ab1{1213,34343},3433,xy1[{3232,989},{8737,
+933}],3343,pq1[{1212,989}],98383},8383,xty)';
 print qq{'$s'};
 ;;
 my $curly_grp  = qr{ { \d+ (?: , \d+)* } }xms;
 my $curly_grps = qr{ $curly_grp (?: , $curly_grp)* }xms;
 my $string = qr{ [[:alpha:]] [[:alnum:]]+ }xms;
 ;;
 $s =~ s{
   $string    \K ($curly_grp)
   |
   $string \[ \K ($curly_grps) (?= \])
   }
   { (my $one = ($1 // $2)) =~ s{,}{|}xmsg;  $one; }xmsge;
 print qq{'$s'};
"
'(12,4666,{xyx,yyy,ab1{1213,34343},3433,xy1[{3232,989},{8737,933}],334
+3,pq1[{1212,989}],98383},8383,xty)'
'(12,4666,{xyx,yyy,ab1{1213|34343},3433,xy1[{3232|989}|{8737|933}],334
+3,pq1[{1212|989}],98383},8383,xty)'
[download]

(Beware possible wraparound!) (Update: Also be aware that this approach only works for one single level of nesting! I think a regex multi-level nesting solution is possible, but perhaps a bit tricky.)

Update: The substitution can be made a bit simpler with the  (?|alternations) extension of 5.10+ (which causes both captures to go to $1), and the  s///r or  tr///r (new /r modifier) extension of 5.14+. E.g. (tested):

$s =~ s{
  (?| $string    \K ($curly_grp)
      |
      $string \[ \K ($curly_grps) (?= \])
      )
  }
  { $1 =~ tr{,}{|}r }xmsge;
[download]

---

Give a man a fish:  <%-(-(-(-<

I tried the below expression. It worked for few cases and failed for other cases. Dont know the reason. First expression is to replace comma between <string>{ and } and other to replace comma between <string>[{ and }]

perl -pe 's/,(?=[^\{\}]*\d\b*\{*\})/|/g;s/,(?=[^\[\]]*\[*\{*\}*\])/|/g
+;'
[download]

You still haven't explained what "failed in other cases" means.

It's hard to read in your one-liner (have you tried adding the /x flag to your regexp which allows you to write it over multiple lines, with comments?), but in your second statement it sure seems like you are replacing anything that matches ,(?=[^\[\]]*\[*\{*\}*\]) with | ... you probably need to learn about extracting matches, too.

Remember: Ne dederis in spiritu molere illegitimi!