Filehandles

You can create your own filehandles using the open function, read and/or write to them, and then clean up using close.

open

open opens a file for reading and/or writing, and associates a filehandle with it. You can choose any name for the filehandle, but the convention is to make it all caps. In the examples, we use FILEHANDLE.

open a file for reading

open FILEHANDLE,"cosmids.fasta"

alternative form:

open FILEHANDLE,"<cosmids.fasta"

open a file for writing

open FILEHANDLE,">cosmids.fasta"

open a file for appending

open FILEHANDLE,">>cosmids.fasta'

open a file for reading and writing

open FILEHANDLE,"+<cosmids.fasta"

Catching Open Failures

It's common for open to fail. Maybe the file doesn't exist, or you don't have permissions to read or create it. Always check open's return value, which is TRUE if the operation succeeded, FALSE otherwise:

$result = open COSMIDS,"cosmids.fasta";
die "Can't open cosmids file: $!\n" unless $result;

When an error occurs, the $! variable holds a descriptive string containing a description of the error, such as "file not found".

There is a compact idiom for accomplishing this in one step:

open COSMIDS,"cosmids.fasta" or die "Can't open cosmids file: $!\n";

Using a Filehandle

Once you've created a filehandle, you can read from it or write to it, just as if it were STDIN or STDOUT. This code reads from file "text.in" and copies lines to "text.out":

open IN,"text.in" or die "Can't open input file: $!\n"; open OUT,">text.out" or die "Can't open output file: $!\n"; while ($line = <IN>) { print OUT $line; }

Here is a more idiomatic way to do the same thing:

while (<IN>) {
   print OUT;
}

And the minimalist solution:

print OUT while <IN>;

Closing a Filehandle

When you are done with a filehandle, you should close it. This will also happen automatically when your program ends, or if you reuse the same filehandle name.

close IN or warn "Errors while closing filehandle: $!";

Some errors, like filesystem full, only occur when you close the filehandle, so you should check for errors in the same way you do when you open a filehandle.

Pipes

You can open pipes to and from other programs using the pipe ("|") symbol:

open PIPEIN,"ls -l |" or die "Can't open pipe in: $!\n"; open PIPEOUT,"| sort" or die "Can't open pipe out: $!\n"; print PIPEOUT while <PIPEIN>

This mysterious example runs the ls -l command, reads its output a line at a time, and does nothing but send the lines to the sort command.

Here's an actually useful example that counts the number of times that the string pBR322 appears in a file. If the file ends with a .gz suffix (using a pattern match, which we talk about later), it opens a pipe from the gunzip program to uncompress the file on-the-fly. Otherwise it does a normal open:

#!/usr/local/bin/perl $file = shift or die "Please provide a file name"; if ($file =~ /\.gz$/) { # a GZIP file open IN,"gunzip -c $file |" or die "Can't open zcat pipe: $!\n"; } else { open IN,$file or die "Can't open file: $!\n"; } $count = 0; while (<IN>) { chomp; $count++ if $_ eq 'pBR322'; } close IN or die "Error while closing: $!\n"; print "Found $count instances\n";

Reopening Standard Input

If you try to open the special filename "-" for reading, you get a filehandle attached to standard input. This is useful for processing a list of filenames on the command line:

(~) 50% process.pl foo.txt - bar.txt

"-" is actually a Unix convention for standard input.

Previous

Contents

Next