If you need to do the substitution on .txt files also, open a different output file and then rename it to where you want it, or read in the entire file before opening it for writing.

thanks a lot . that was exactly the problem . it was working fine on other extensions but not on txt files .

On unix it would seem that you are truncating the text files by opening them to write before you have read them, a result of opening them by the same name. I don't recall how that works on win32, but I suspect it is the same.

You can fix that by either using a temporary file name and renaming it after the copy is done, or else by opening in '+<' mode, which opens to read and write, without truncation.

Here's a rewrite using the non-truncating r/w open. I prefer it to the temporary file because of race conditions which may occur with temporary files. I'll use glob to simplify grabbing the file listing.

#!/usr/bin/perl
use warnings;
use strict;
use Fcntl ':flock';

my $path = 'c:/perl/dvd_files/';

# only doing the problem files - those with unchanged names
my @files = glob "$path*.txt"; 

for (@files) {
    local $/;
    open my $fh, '+<', $_ or warn($!), next;
    flock $fh, LOCK_EX;
    binmode $fh;
    my $text = <$fh>;
[download]

    $text =~ tr/0-9{}//d;
    seek $fh, 0, 0;
    truncate $fh, 0;
    print $fh $text;
}
[download]

On reflection, I wonder if you wouldn't be better off creating the rewritten files in a brand-new subdirectory and not clobbering anything. That may be necessary, anyway, since I can't test on win32 and am not sure that flock, seek, or truncate will work on your system.

thanks zaxo , i agree with you that i overdid somethings. your code seems to work much better for me . i have another query though . this code that i am posting now is taken from a site for stripping HTML . It does the same thing that my other code was doing , generated empty txt files and gives me "Use of uninitialised variable " error for $plain_text. can any body tell me why??

#!c:/perl/perl.exe
use strict;
use warnings;
use CGI ':standard';
use HTML::Parser;

my $plain_text ;
my $p = HTML::Parser->new(text_h => [\&text_rtn, 'text']);


my $path = "c:/perl/htmlfiles";
opendir(LOCAT, $path) or die "Couldn't open folder, $!\n";
my @folder = grep !/^\.\.?$/, readdir(LOCAT);
closedir (LOCAT);

# Do a loop for each file in the folder. This gets the filename also.
foreach my $file (@folder)
{
    my $full_path = $path.$file;
    print "Reading '$full_path'\n";
    $p->parse_file($full_path);
    $full_path =~ s/html/txt/gi;
    $full_path =~ s/htm/txt/gi;
    print "Writing '$full_path'\n";
    open(WRITE,">$full_path") or die("Cannot create file!");
    print WRITE $plain_text;
    close(WRITE);
}

sub text_rtn {
    foreach (@_) {
    $plain_text .= "$_\n";
    }
}
[download]

You might want to reconsider the substitutions that operate on $full_path. As they are now, ANY occurance of 'sub' or 'srt' in $full_path will be replaced with 'txt', not just the file extension (which I assume you are trying to do). If you want to replace only the extension, you could do something like:

$full_path =~ s/sub$/txt/i;
[download]

Alternatively, you could use the File::Basename and/or File::Spec modules to perform the path/filename operations for you (i.e., split off the suffix and create a path/filename for the new file).

perl -pi.bak -e 's/[\{0-9\}\{0-9\}]//g;'
[download]