push $item, $_ if /$pattern/;

Pushing to a scalar (which isn't an array ref) is an error. Just use an array instead. Same goes for your foreach postfix. Don't declare $item inside and outside the loop unless you intend them to be different variables. And maybe think about an alternative style of indenting? Good luck.

for (@bpout) {
    print "$_\n"  if /$pattern/;
}
[download]

Fantastic information. I was wondering though, you don't seem to be able to use the "filehandle" variable like a real file in a script. For example: I have this beginnerish script

open my $fh2, '<', '/tmp/mlist.txt' or die "unable to open file '$file
+' for reading : $!";
open my $fh9, '>', '/tmp/listk.txt' or die "unable to open file '$file
+' for reading : $!";
while (<$fh2>) {
   my @fields = split(',', $_);
 local $" = ',';
       if ($fields[2] eq '1' || $fields[2] eq '0') {
             print $fh9 "@fields[4]\n" if /MONTHLY/ && !/,-,/;
             }
           }
           close $fh2;
           close $fh9;
[download]

This is clumsy but it works. However, if I try to use the variable assigned to a filehandle in the same way:

open my $bpfh1, '<', \$forbpjoutput or die $!;
  while (<$bpfh1>) {
my @fields = split(',', $_);
my $fields;
 local $" = ',';

       if ($fields[2] eq '1' ||$fields[2] eq '0') {
             print "THESE ARE THE>>> $fields[4]\n" if /MONTHLY/ && !/,
+-,/;
        
close $bpfh1;
             }
          }
[download]

it does work but what happens is that it never stops printing the "file" or the fields4 It continues until out of memory. Is there a way to do this differently? This is probably a very basic question, but I don't have much experience with filehandles which are variables.

Hi novice2015

The link to open that BillKSmith mentioned in his post has the following text.

# in-memory files
21. open(my $memory, ">", \$var)
22.     or die "Can't open memory file: $!";
23. print $memory "foo!\n";              # output will appear in $var
[download]

This shows how to write to an in-memory file.

To read from an in memory file, you would do this:

#!/usr/bin/perl
use strict;
use warnings;

my $file = 'here
are
some
words
I
want
';

open my $fh, '<', \$file;

chomp(my @words = <$fh>);

my $pattern = join "|", @words;

print $pattern;
[download]

That would be a way to not need the tmp. file.

I cannot duplicate your problem. When I run your code on my system (perl v5.20.2 on Windows 10), it terminates with no error messages and no output. It I use warnings, I get several copies of a warning about uninitialized value in $bpfh1. If I declare $forbpjoutput as a lexical variable and define it as a null string, I get no messages and no output even using strict and warnings. (This is exactly what I expect).

use strict;
use warnings;
my $forbpjoutput = q();
open my $bpfh1, '<', \$forbpjoutput or die $!;
... # Exactly as in your post.
[download]

As Cristoforo pointed out, you probably did not intend to input from a null file. I cannot guess what you really did intend to do.

Bill

Hi novice2015,

I don't think in-memory files like you are trying to use here are necessary, they overcomplicate the solution. Just use a normal Perl array to hold the data in between loops.

What is unclear to me is this: If you don't want to use any of the files in /tmp, where will the data that you are currently reading from /tmp/bpl.txt and /tmp/lst.txt come from? In order to answer your original question, I could assume the data you are reading from those two files is available in @words and @data, respectively. Then I can answer the original question and rewrite your original script like so:

# get @words and @data from somewhere
my $pattern = join '|', @words;
my @lines;
for (@data) {
    push @lines, $_ if /$pattern/;
}
# do something with @lines
[download]

By the way, are you using warnings and strict? In your original code you wrote open my $bpl_fh, ... but then close $bplist_fh;. Also, in your later posts it looks like you're closing $bpfh1; too soon, inside the loop, which should give you a warning. So always Use strict and warnings!

Now, if I understand your later posts, you then want to take the lines that you've previously matched with /$pattern/ and stored in @lines and operate on those. If I take your code from here and rewrite it, I get:

for (@lines) {
    my @fields = split /,/;
    if ( $fields[2] eq '1' || $fields[2] eq '0' ) {
        print "THESE ARE THE>>$fields[4]\n" if /MONTHLY/ && !/,-,/;
    }
}
[download]

But then my question is, do you need @lines at all? You could skip the creation of that array and do:

my $pattern = join '|', @words;
for (@data) {
    next unless /$pattern/;
    my @fields = split /,/;
    if ( $fields[2] eq '1' || $fields[2] eq '0' ) {
        print "THESE ARE THE>>$fields[4]\n" if /MONTHLY/ && !/,-,/;
    }
}
[download]

If you have further questions, I strongly recommend you read Short, Self Contained, Correct Example and How do I post a question effectively? - without input data, runnable code, expected output or exact error messages, it's a guessing game.

Hope this helps,
-- Hauke D