• From a style and a code maintenance point of view, it might be better to move the closedir(TEMP); line up to immediately after the my @FILES=readdir(TEMP); line since you've done with reading the directory handle at that point.
• It is now considered good practice to use the three-argument form of open with lexical filehandles and, when testing for success as you do (well done ++), it is useful to output the O/S error $! (see perlvar).
• It is perhaps better to use Perl-style for loops rather than C-style ones. You can use for and foreach interchangeably. Instead of

  my $fi = scalar (@FILES);
  for (my $k = 0; $k < $fi ; $k++)
    {
       open (F,$FILES[$k]) or die " rey ledhu ra \n";
       ...
    };
  [download]

  use

  foreach my $file ( @FILES )
  {
      open my $inFH, '<', $file
         or die "open: < $file: $!\n";
      ...
  }
  [download]

Cheers,

JohnGG

I would actually suggest forgoing the loop altogether and using grep:

my @FILES = grep { -f } readdir (TEMP);
foreach my $file (@FILES){
  open (F, $file) or warn "Couldn't open $file: $!" and next;
  my @matches = grep {  /$str/ }  <F>;
  close (F);
  next unless scalar @matches > 0;
  print join("\n", @matches); 
}
[download]

Of course, that prints every matching line in the file, if you only want the first (evidenced by "last"), you could just use:

print $matches[0], "\n";
[download]

And do you really want to use "die" in a loop? What if you can't open one file, but you can open the rest? Isn't "warn and next" more appropriate?

I would actually suggest forgoing the loop altogether and using grep

Nothing wrong with using grep to fill the @FILES array, although I might do grep { -f and not -l } depending on how symlinks should be treated. However, unless my eyes are failing me, you still have the loop in your code.

And do you really want to use "die" in a loop?

The OP seems to and, since I don't know what the requirements are, I followed his/her lead.

Cheers,

JohnGG

I'm afraid I don't really understand your question. You talk about "admin" and "admin/1.1", but in your example code (++nave448 for providing it), the search string is "string to be matched". What are you actually looking for here?

Hi,
thanks for the reply
"string to be matched" is the search string
for example if i give it as admin/1.1 it gives me the file that contains the string admin/1.1
But if i give it as admin it also gives me the same file
I just want to match the whole word admin/1.1 not the part admin
regards, Naveen

No need for a module, the quotemeta built-in function does that. You can also use \Q$searchstring\E in the regex (which does the same).

Thanks. I got the problem

Check out the perlre man page, which explains in good detail how to work with regex.

When using characters that are normally interpreted as syntax, in your case (/ and .) these characters must be escaped. To have a slash (/) represent a slash literally, you designate the character to be escaped by adding a backslash in front of it like so (\/). This tells perl to interpret the slash / literally. The same is true of the period (\.) and other characters that are normally used as perl syntax.

Hope this helps!