Open File in Perl

After explaining how to open a file for writing and how to read from a file in Perl I got a lot of comments that I should teach more "modern ways" for handling files. People pointed to IO::All and that can certainly be intersting but I found File::Open by Lukas Mai to be nice and simple. I even extended it and created File::Open::OOP.

There are various ways to read from text a file in Perl. We have already seen how to do it with the tools available in the language itself. Now we are going to see an even simpler way using a modules from CPAN.

Open a text file for reading

If you'd like to read from a text file line-by-line, first you need to open it.


  use File::Open qw(fopen);
  my $fh = fopen $file;

  while (my $row = <$fh>) {
     print $row;
  }

In this example we are using the File::Open module written by Lukas Mai which is a thin wrapper around the open function of Perl. It provides an interface that, to me, looks much nicer than the standar open function.

There are 2.5 differences:

fopen returns the file-handle instead of putting it in the first parameter.

It throws an exception when the file cannot be opened so it eliminates the need of the or die part and that of autodie for open.

It defaults to reading a file but you can supply the opening mode as a parameter.

Open file for writing

If you'd like to open a file for writing you would write:


  fopen $file, '>';

or better yet:


  fopen $file, 'w';

That's the half difference. That you can use letters to indicate the file opening mode.

Reading non-ASCII files

When reading a file that is saved with UTF-8 encoding it is important to explicitelly tell this to Perl. The way to do that with fopen is to pass the "layers" to the function:


  my $fh = fopen $file, 'r', 'utf8';

This way when we read from the file the data will be assumed to be in UTF-8 encoding.

Reading files in Object Oriented way

Warning: This is now experimental!

I thought it might be nicer to provide an OOP interface to the reading and writing part so I started to extend the File::Open module and created File::Open::OOP.

Currently it works like this:


  use File::Open::OOP qw(oopen);

  my $fh = oopen 'filename';
  while ( my $row = $fh->readline ) {
    print $row;
  }


oopen accepts parameters just as fopen does but instead of returning a file-handle it returns a File::Open::OOP obect. That object can be then used to read the content of the file line-by-line using its readline method.

I am not sure if this is nicer but then it can also be used in two other ways:

The readall method will return all the rows as a list.


  my @rows = oopen('filename')->readall;

  foreach my $row (@rows) {
  }


The slurp method will read all the lines into a single scalar:


  my $rows = oopen('filename')->slurp;

This is not dependent on the environment. Even if I assign it to an array it will always return a single string.


  my @rows = oopen('filename')->slurp;

What do you think?

Does this have any advantage? Should one just use File::Slurp instead?

Should this oopen function be actually a part of the File::Open module? I sent an e-mail to Lukas Mai but I have not received a reply yet. Maybe you know how to find him?


Perl tutorial

For further articles see the Perl tutorial.

Published on 2011-12-21 by Gabor Szabo

In the comments, please wrap your code snippets within <pre> </pre> tags and use spaces for indentation.
blog comments powered by Disqus
EduMaven
Python, Dart, JavaScript, and more.


Google Plus Twitter RSS feed