Why don't you use JSON to create your output. Instead of struggling with complicated strings, you would only have to create some hashes and let the module do the work for you. For example:

use strict;
use warnings;
use JSON;

my %bookmarks;
for (1..3) {
        push @{$bookmarks{browserBookMarksData}}, { bookmark => "0", c
+reated => "01/12/2012" };
}
print to_json \%bookmarks;
[download]

would create

{"browserBookMarksData":[{"created":"01/12/2012","bookmark":"0"},{"cre
+ated":"01/12/2012","bookmark":"0"},{"created":"01/12/2012","bookmark"
+:"0"}]}
[download]

No need to worry about commata...

Update: The last part of your script could look similar to this:

print "Enter the limit : ";
chomp(my $lim = <STDIN>);

my %bookmarks;
while ( $lim-- ) {
    push @{$bookmarks{browserBookMarksData}}, {
        bookMark => $charset[ rand(@charset) ],
        created  => $charset2[ rand(@charset2) ],
        date     => $charset3[ rand(@charset3) ],
        title    => join '', @charset4[map {int rand @charset4} (1..8)
+],
        url      => join '', @charset5[map {int rand @charset5} (1..8)
+],
        visits   => $charset6[rand(@charset6) ],
    };
}
print to_json \%bookmarks;
[download]

Hi , Thanks for the replies. I updated my json which is as follows--

use strict;
use warnings;
use JSON;
open(FILE, ">>bookmarkmirror.json");
my %bookmarks;
my @charset =  ('1','0');
my @charset2 =('');
my @charset3 =('');
my @charset4 = (('A'..'Z'), ('a'..'z'));
my @charset5 = (('A'..'Z'), ('a'..'z'));
my @charset6 = ('1' .. '1000');
my @charset7 =('http://www.','https://www.');
my @charset8=('.com','.co','.in','.info','.org','.net','.biz','.us','.
+me','.mobi','.co.in','.firm.in','.gen.in','.ind.in','.net.in','.org.i
+n','.tv','.ag','.am','.asia','.at','.be','.bz','.ca','.cc','.co.nz','
+.co.uk','.com.ag','.com.au','.com.br','.com.bz','.com.co','.com.es','
+.com.mx','.com.pe','.com.so','.com.tw','.cz','.de','.es','.eu','.fm',
+'.fr','.gs','.idv.tw','.it','.jobs','.jp','.la','.me.uk','.ms','.mx',
+'.net.ag','.net.au','.net.br','.net.bz','.net.co','.net.nz','.net.pe'
+,'.net.so','.nl','.nom.co','.nom.es','.nom.pe','.org.ag','.org.au','.
+org.es','.org.nz','.org.pe','.org.so','.org.tw','.org.uk','.pe','.se'
+,'.so','.tk','.tw','.ws','.xxx');


my $i=0;
print "Enter the limit : ";
chomp(my $lim = <STDIN>);


while ( $lim-- ) {
    push @{$bookmarks{browserBookMarksData}}, {
        bookMark => $charset[ rand(@charset) ],
        created  => $charset2[ rand(@charset2) ],
        date     => $charset3[ rand(@charset3) ],
        title    => join '', @charset4[map {int rand @charset4} (1..8)
+],
        url      => join '', @charset5[map {int rand @charset5} (1..8)
+],
        visits   => $charset6[rand(@charset6) ],
    };
}
print FILE to_json \%bookmarks;
  close(FILE);
[download]

The ouput is pasted below-

{"browserBookMarksData":{"created":"","bookMark":"1","date":"","title":"YvJzwzBRurlOgWzGkJEvisits843"},{"created":"","bookMark":"1","date":"","title":"COubQkTRurlfVbzdDGUvisits393"}}

Now the problem is that the url and visit fields get merged .There is no comma seperation between there the url and the visits fields and also how to join the three character set ie. charset4 ,charset7 and charset 8 to make the url field.

In order to avoid join to consume the rest of the list, try

        title    => join( '', @charset4[map {int rand @charset4} (1..8
+)]),
        url      => join( '', @charset5[map {int rand @charset5} (1..8
+)]),
[download]

Examine your output to narrow down the line of code that adds the unwanted commas

Simplify/neaten your output so you can notice which comma is the unwanted one

ddumper^{Basic debugging checklist} Basic debugging checklist item 4, and perltidy^{Basic debugging checklist} Basic debugging checklist item 10

Your program with perltidy -olq -csc -csci=10 -cscl="sub : BEGIN END if " -otr -opr -ce -nibc -i=4 -pt=0 "-nsak=*"

While you're at it you might fix the syntax errors (if inside print)

And get rid of the extra curly braces ... and the seperation of declaration and initialization

            my $bookMark = $charset[ rand( @charset ) ];
[download]

And the back slashing -- the battle of Hastings was very long ago, use alternate delimiters

    print FILE ( qq{
      {
         "bookMark": "$bookMark",
         "created":  "$rand1",
         "date":     "$rand2",
         "title":    "$title",
         "url":      "$www$url$com",
         "visits":   "$visits"
      }
} );
[download]

STDIN is a non-fun way to pass a single number, use @ARGV with a default

my $lim =  int( shift || 3 );
warn "# The limit is $lim\n";
[download]

Don't make variables out of things that aren't variable, get rid of $comma/$prefix/$sufix/$curly

This ought to get you 99% there leaving only the comma logic which ought to be more apparent now

JSON parsers don't care if the output is tidy, humans however DO

jsontidy

#!/usr/bin/perl --
use Path::Tiny;
use JSON;
Main( @ARGV );
exit( 0 );
sub Main {
    my $in = shift or die "\nUsage: $0  foo.json  > foo-tidy.json\n";
    
    binmode STDOUT, ':encoding(UTF-8)';
    print JSON->new->utf8(1)->pretty(1)->encode(
        JSON->new->decode(
            path( $in )->slurp
        )
    );
}
[download]

Is this JSON producer a learning exercise, or is it to create a JSON encoder that does something the others don't? Is the issue that you're not allowed to install 3rd party modules? I'm just curious. There may be a more robust solution that consumes less work than the path you're currently on. Even if your specification is to not use any modules from CPAN, modules such as JSON::Tiny are set up to allow copy/paste embedding if your situation requires it.