If you just want to replace a word, you don't even need a script:

perl -pi.orig -e 's/original/substitute/g' <filename>

Each time you read something from <FILEHANDLE> in scalar context it reads one line.

Perhaps you should find a good book to learn Perl from.

perl -pi.orig -e 's/original/substitute/g' <filename>

(To the OP:) if you want that to apply to a large number of files, all in the same directory, just give them on the cmd line. That is, <filename> can really be <filenames> or a shell pattern that the shell will expand. If the files are too many for the shell to handle, then as I wrote in my other reply do so yourself with glob:

perl -pi.orig -e 'BEGIN{@ARGV=map glob, @ARGV} s/original/substitute/g
+' '*.foo' 'bar.*'
[download]

Note that this is also necessary under Windows, where the shell does not do expansions at all. But in that case you have to use double quotes instead of single ones.

hmmm .. ok .. but if i want that it will search, check and replace all words in all files in a folder? look, i have for example about 1000 TXT files, and there is a word Bartholome and i need to replace it for Richard - In ALL files in this folder ... Yes i know .. I do perl mostly for fun, but i need it now .. :(

look, i have for example about 1000 TXT files, and there is a word Bartholome and i need to replace it for Richard

Similar to what was already pointed out to you:

perl -spi -e 's/Bartholome/Richard/g' /path/to/folder/*
[download]

---

Peter L. Berghold -- Unix Professional
Peter -at- Berghold -dot- Net; AOL IM redcowdawg Yahoo IM: blue_cowdawg

hmmm .. ok .. but if i want that it will search, check and replace all words in all files in a folder?

Use glob or File::Find (if you want to do that recursively) or one of its friendlier relatives.

Could you please help me solve this two problems? Here is my source:

Far too many problems to deal with all of them in an amount of time that could be reasonable for me now. Just some hints:

• "PERL";
• no strict - too bad!
• weird style;
• "old style", unchecked opens;
• no proper iteration over the lines of a file nor slurping nor... anything!

Please do not take it as a personal offense, but it seems to me that you're trying to do something in Perl without even having a clue about the very basics of the language. All in all you may well be blindingly gluing together fragments found who know where. I recommend you get an introductory book or tutorial instead.

Perl (not PERL) does not, usually, loop through all the lines of the file; to make it do so see perlrun.

I've a few questions about your post:

• Do you really want to replace the word in binary files? Changing "GIF" to "IMG" may make a binary file unusable. Use a file test, such as -T to see if the file is text or not (note that I've found -T and -B to be less than 100% reliable on Windows, so I would be careful on any platform).
• If you want to change every file in a directory, you will have to loop through the files, not just loop through the records in a file. My preferred way of doing this is glob to get the file names, and looping through the resulting list.
• You can change files in place; there is a Perl runtime option to do so. See perlrun. I believe the option is -i. I tend not to use this option; when I need to change a file in situ my preference is Tie::File.
• Just about every language with which I've dealt is case sensitive when dealing with strings; in Perl (and other languages) this can be turned off in regular expressions (regex); see perlretut
• Depending on the size of your files, you may want to localize $/ (see perlvar) so each file's contents are read as a single string, vs reading the file line by line, e.g., something like this:
  

  use strict;
  use warnings;
  
  sub subsitute {
    (my $filename, my $oldword, my $newword) = @_[0..2];
  
    local $/;
  
    if(open(my $fh, "<", $filename)){
       my $contents = <$fh>;
  
       $contents =~ s/\b$oldword\b/$newword/ig;
  
       close($fh);
  
       if(open($fh, ">", $filename){
          print $fh $contents;
          close($fh);
          return 1; # for success
          }
       else{
          warn "Could not open $filename for writing because $!; skippin
  +g it\n";
          return undef;
          }  
     }
    else {
     warn "Could not open $filename because $!; it will be skipped\n";
     return undef;
   }
  }
  [download]

  I'm explicitly opening the file for writing as I don't want to assume that the -i runtime option is specified. Also, I'm assuming that case is not important and that you want every instance of $oldword replaced.

You will have to do some more checks than just to see if the file is binary or not. Obviously, you should make sure that you've got correct rights to change the file. I would also advise that you copy the file, instead of changing it in place, until after you've thoroughly tested the code. You would, for example, not want to change source code in such a way as to introduce a syntax error or undefined external reference. And changing system("cp $oldfile $newfile"); to system("rm $oldfile $newfile"); could tend to be career inhibiting.

Hi, thankx for the help, BUT nothin of this is working ...

I tryed all the examples which were mentioned here ...

perl -pi.orig -e 's/original/substitute/g' <filename>
[download]



or

perl -spi -e 's/Bartholome/Richard/g' /path/to/folder/*
[download]



The only thing what i achieved was, that my command line told me that i cant work with a file without and backoup or something like that ...

Than even the full script, but it was full of errors, about 2-3 ...

Guys, i know that i am like an insect but i really need help ... And you are my only hope - or i have to spend hours replacing a word by hand ... Please write me an exact way how to do that ...

Than, if youll be in Slowakia lemme know, ill buy you a beer ... Thank you

perl -p -i_ARCHIVE.txt -e "s/original/subst/g"  <path to FILE>
[download]



good, afrer HOURS spend on googling the net i found this one ... Its workning on MS WIN without problems ...

It replaces the word, saves it under original form {name and file type} and creates a backup in form : original_ARCHIVE.txt ...

Nice ... And now the last thing what i need is, how to work with more files? If i type in as path to file:



C:/test/*     - Does not work
C:/test/*.*   - Does not work
C:/test/"*"     - Does not work
C:/test/"*.*"   - Does not work
[download]



And so i have NO idea hot to let him work with all files in a folder, what you told me does not work ..