First, change your substitution regexp to: s/$pattern/' ' x length( $1 )/gie; ...note the use of $1 instead of $&. It is advisable to always avoid $& if possible for performance reasons.

Now to answer your question: Use a second regexp to keep it simple, and it should look like this:

$target_data =~ s/(\n{2,})/'\n' . ( ' ' x length( $1 ) ) /ge;
[download]

That converts any blank lines to a line with spaces in it equalling the number of blank lines found in a row. If you don't want the blank spaces to occupy a new line, remove the '\n' part from the right side of the substitution regexp.

This is not what I want. I want is to change the lines if any in that pattern found, not in the whole data. The sample data is

</TABLE>

2
<PAGE>

abcdefghijklmnopq
[download]

I hope that u understand my problem.

davido++ has given you a solution which replaces every newline with a space. Isn't that exactly what you're asking for?

If not, you'll need to show several examples - enough that we can see what you want. Make sure the spaces are countable.

After Compline,
Zaxo

Update: Finally grasped the feature that the OP wanted and wasn't getting: In addition to substituting spaces, maintain the same number of lines. Like this:

$_ =<<'EOS';
</TABLE>

2
<PAGE>

RISK/RETURN SUMMARY AND FUND EXPENSES


PRIME MONEY MARKET FUND
EOS

my $pat = qr/(\n\s*[0-9]{1,3}\s*\n\s*<page>\s*\n\s*)(.*\n)/i;

s{$pat}{
  my ($m1, $m2) = ($1, $2);
  my $num_lines = $m1 =~ tr/\n//; # Just counting lines
  $m1 =~ tr//!/c; # Replacing all chars with ! (for visibility; use sp
+ace instead in final code)
  $m1 . $m2 . "\n" x $num_lines;
}gie or warn "No match!\n";

print;
[download]

</TABLE>!!!!!!!!!!!!RISK/RETURN SUMMARY AND FUND EXPENSES







PRIME MONEY MARKET FUND
[download]

my $pattern='(</TABLE>\n\s*[0-9]{1,3}\s*\n\s*<page>\s*)(.*)(\n*)';
$target_data =~ s/$pattern/$2." " x length($3)/gie;
[download]

The sample data is

</TABLE>

2
<PAGE>

abcdefghijklmnopq
[download]

I hope that u understand my problem. Please help me to solve the problem.

And this should become... ?

</TABLE>

2
<PAGE>

RISK/RETURN SUMMARY AND FUND EXPENSES


PRIME MONEY MARKET FUND
[download]

</TABLE>            RISK/RETURN SUMMARY AND FUND EXPENSES







PRIME MONEY MARKET FUND
[download]

Edit by castaway - reparented under original question

Hi,

Is this solving your problem. It solves for the sample data you have posted



use strict;

my $target_data="</TABLE>\n\n2\n<PAGE>\n\nRISK/RETURN SUMMARY AND FUND
+ EXPENSES\n\n\nPRIME MONEY MARKET FUND";

my $pattern='(\n\s*[0-9]{1,3}\s*\n\s*<page>\s*\n)(.*)(\n*)';
$target_data =~ s/$pattern/" " x length($1).$2."\n" x length($3)."\n" 
+x length($1)/gie;

print $target_data;

print "\n\n";
[download]



The output is



</TABLE>            RISK/RETURN SUMMARY AND FUND EXPENSES














PRIME MONEY MARKET FUND
[download]

If the "\n" looks too many then you can change my regx concatenation of the regx to minimize it

"\n" x length($3)."\n" x length($1)/gie;
[download]

Thanks
SasiKumar

Thanks Sasi but the above code doesn't solve my problem. As it is not the general problem for the all patterns found but it is just for the above example. As I had told earlier that the location of the pattern or the no. of lines in the pattern found is not fixed and there is no surity that the page no found could be only that way. The pattern found could be

abc
The pattern follows 
<page> 24 

def
[download]

Like for the above case the code will not work. I had tried this code before also. I had made the pattern for the above cases as

$pattern='(\n\s*[0-9]{1,3}\s*\n\s*<page>\s*\n)';
if ($target_data =~ /$pattern/gi)
   {
$target_data =~ s/$pattern/" " x length($1)." " x length($2)." " x len
+gth($3)." " x length($4)/gie;
}
$pattern='(\n\s*-[0-9]{1,3}-\s*\n\s*<page>\s*\n)';
if ($target_data =~ /$pattern/gi)
    {
    $target_data =~ s/$pattern/" " x length($1)/gie;
}
$pattern='(\n\s*[A-Za-z]-[0-9]{1,3}\s*\n\s*<page>\s*\n)';
if ($target_data =~ /$pattern/gi)
    {
       $target_data =~ s/$pattern/" " x length($1)/gie;
     }
$pattern='(\n\s*page\s*[0-9]{1,3}\s*of\s*[0-9]{1,3}\s*\n)';
if ($target_data =~ /$pattern/gi)
    {
       $target_data =~ s/$pattern/" " x length($1)/gie;
    }
[download]

But the above patterns are just lacking in the counting the no. of lines if found in the data, that should be converted into spaces. I hope that now u have a clear picture of the problem.