Re^2: Shorten the headers of a file and remove empty lines using perl

How can I input the "input_file"? "$input_file" is a scalar but not a file. I run the following script:

#!/usr/local/bin/perl
use warnings;
open ( my $input_fh, "<", $genome );
open ( my $output_fh, ">", $output_file );
foreach my $line ( <$input_fh> )
{
  unless ( $line =~ m/\A\s*\Z/ )
  {
    $line =~ s/(GL\d{6})\d+/$1/;
    print $output_fh $line;
  }
}
close ( $input_fh );
close ( $output_fh );
[download]

and got the following messages:

Name "main::genome" used only once: possible typo at header.pl line 3. Name "main::output_file" used only once: possible typo at header.pl line 4. Use of uninitialized value $genome in open at header.pl line 3. Use of uninitialized value $output_file in open at header.pl line 4. readline() on closed filehandle $input_fh at header.pl line 5.

I think my way to open the file is wrong. Can you give me some hints? Thanks XF

Comment on Re^2: Shorten the headers of a file and remove empty lines using perl Download Code

Replies are listed 'Best First'.
Re^3: Shorten the headers of a file and remove empty lines using perl by Cristoforo (Curate) on Jun 13, 2013 at 23:27 UTC
`foreach my $line ( <$input_fh> )` should be `while (my $line = <$input_fh> )` The first form is a glob, (but I don't know well enough to explain it to you). The second line should work properly for reading your file. Update: Yes Choroba is correct, it is not a glob - my mistake,	[reply] [d/l] [select]
Re^4: Shorten the headers of a file and remove empty lines using perl by choroba (Cardinal) on Jun 13, 2013 at 23:30 UTC
It is not a glob. It is a readline in list context. If the file is large, it can eat lots of memory, and should therefor be avoided. لսႽ† ᥲᥒ⚪⟊Ⴙᘓᖇ Ꮅᘓᖇ⎱ Ⴙᥲ𝇋ƙᘓᖇ	[reply]
Re^4: Shorten the headers of a file and remove empty lines using perl by Preceptor (Deacon) on Jun 14, 2013 at 20:29 UTC
Indeed - the bit I posted initially will work, but it'll trigger reading the whole file into memory. Probably not a good idea with a 500MB file. Perl contexts are incredibly clever, but do lead to some interesting gotchas - use of a filehandle in a scalar or array context is one of them.	[reply]
Re^3: Shorten the headers of a file and remove empty lines using perl by Cristoforo (Curate) on Jun 13, 2013 at 23:43 UTC
Ah, I didn't see your problem clearly. First, you should `use strict;` as well as `use warnings;`, which you did, in the header of your program. Then, you have to assign the name of your file to `$genome`. `my $genome = 'whateverthename';` (You must assign a name to your output file also)	[reply] [d/l] [select]
Re^4: Shorten the headers of a file and remove empty lines using perl by Anonymous Monk on Jun 14, 2013 at 01:27 UTC
It works very well. Although I do not quite understand all the script yet, I will work on that. Thank all people here for your help!	[reply]
Re^5: Shorten the headers of a file and remove empty lines using perl by Preceptor (Deacon) on Jun 14, 2013 at 20:26 UTC
I skipped the bit where you set up the names of input file and output file. But generally speaking, you should always always 'use strict;' and 'use warnings;'. They really are the very best ways to stop a program doing anything weird. And as mentioned above- a while look is better than a foreach if you're processing a large file. (Makes little odds for a small file, but it's good form). Perl is very clever - it understand context. <$input_fh> says 'read from $input_fh' but if you do: `my $line = <$input_fh>;` [download] it simply reads the next line. Where if you do `my @whole_file = <$input_fh>;` [download] It will read the whole file into that array - which is in effect what my first snippet does. It doesn't make much difference if you're working with a small file, but the difference will become very important with a 500MB file. I'd strongly suggest taking time to understand what each line is doing - code that someone on the internet gave you is never trustworthy. (Although on Perlmonks, usually true evil will get stomped upon)	[reply] [d/l] [select]