in reply to how to save patterns found in array to separate files

Hello 2015_newbie, and welcome to the Monastery!

Since Perl v5.6, you can use lexical filehandles, which behave like other scalar variables. In this case, store the filehandles in a hash with each filehandle keyed to the corresponding filename base. The following code should give you the idea:

#! perl use strict; use warnings; my @carslist = qw(DATSUN_BLUE TOYOTA_MAROON ELANTRA_YELLOW); my %cars; for (@carslist) { unless (exists $cars{$_}) { my $file = 'tmp/' . $_ . '.txt'; open(my $fh, '>', $file) or die "Cannot open file '$file' for writing: $!"; $cars{$_} = $fh; } } my $pattern = join '|', @carslist; $pattern = qr/$pattern/; my @strings; while (<DATA>) { chomp; my @fields = split; if (($fields[1] eq '1' || $fields[1] eq '0') && /$pattern/) { push @strings, $_; print { $cars{$fields[4]} } $fields[0], ',', $fields[4], "\n" if /engineer/ && !/,-,/; } } for (@carslist) { close $cars{$_} or die "Cannot close file '$_': $!"; } print "Strings:\n"; print "$_\n" for @strings; __DATA__ 1250 1 engineer office DATSUN_BLUE office 67 page30 text 1500 1 billing office MERCEDES_TAN office 98 page 40 txt 1500 7 billing office JAGUAR_BLACK office 98 page 40 txt 1250 0 engineer office ELANTRA_YELLOW office 66 page50 txt 1250 1 engineer office DATSUN_BLUE office 67 page30 text

Output:

16:14 >perl 1499_SoPW.pl Strings: 1250 1 engineer office DATSUN_BLUE office 67 page30 text 1250 0 engineer office ELANTRA_YELLOW office 66 page50 txt 1250 1 engineer office DATSUN_BLUE office 67 page30 text 16:14 >

and the files DATSUN_BLUE.txt, TOYOTA_MAROON.txt, and ELANTRA_YELLOW.txt created and populated as required.

Updates: (1) Note also that your script splits on commas, but the data you show is separated by whitespace. I’ve changed the split statement to just my @fields = split;, which is special-cased to split the default variable, $_, on whitespace (as though the pattern were /\s+/).

(2) The braces are required around the filehandle expression in the line:

print { $cars{$fields[4]} } $fields[0], ',', $fields[4], "\n" if /engi +neer/ && !/,-,/; # ^ ^

because “whenever you’re using any expression more complex than a bareword handle or a plain, unsubscripted scalar variable ... you will have to use a block returning the filehandle value...” (perldoc print).

Hope that helps,

Athanasius <°(((><contra mundum Iustus alius egestas vitae, eros Piratica,

Replies are listed 'Best First'.
Re^2: how to save patterns found in array to separate files
by 2015_newbie (Novice) on Dec 29, 2015 at 05:38 UTC
    Greetings Monk!! This works just incredibly fantastic! I was able to adopt the logic to other reports as well. This is my first experience with the monks -they're incredible. I have just one question, can you explain why the following works: $cars{$fields4} } $fields[0], ',', $fields4, "\n" The part I don't understand is why the need for {$cars{$fields[4]}} If I put other numbers than 4, I get compile errors. Why not just  $cars{$fields[0,4]} } This is probably a silly beginner question but I don't have much perl experience and am wondering why the "4" is necessary there.

      Hello again, 2015_newbie,

      A typical print statement has the following syntax:

      print FILEHANDLE LIST

      where LIST is a list of what things to print, and FILEHANDLE is a special value telling Perl where those things are to be printed. If FILEHANDLE is omitted, the destination defaults to the last selected filehandle — typically STDOUT, which means print to the screen. If you want your output to go to a file — as you do here — then you need to open the file for writing and associate that opened file with a filehandle variable; then you give that filehandle to print to tell it where to print to.

      Now, let’s return to the line:

      print { $cars{$fields[4]} } $fields[0], ',', $fields[4], "\n" if /engineer/ && !/,-,/;

      It might make things clearer if we add a couple of variables within the if clause:

      if (($fields[1] eq '1' || $fields[1] eq '0') && /$pattern/) { push @strings, $_; my $number = $fields[0]; my $model = $fields[4]; print { $cars{$model} } $number, ',', $model, "\n" if /engineer/ && !/,-,/; }

      Now it’s easier to see: in the LIST, the variable $model is what gets printed (second); but in the expression which denotes the filehandle, $model is the key corresponding to the filehandle stored in the hash %cars. So the value of $model (i.e., $fields[4]) might be DATSUN_BLUE (a string), but the corresponding hash entry, $cars{$model} (i.e., $cars{$fields[4]}), will be a special value — one which Perl understands — which identifies the file “tmp/DATSUN_BLUE.txt” as the file for print to write to.

      It might help you to study perldata to get a good grasp of arrays and hashes.

      Update: Corrected typo and improved wording.

      Hope that helps,

      Athanasius <°(((><contra mundum Iustus alius egestas vitae, eros Piratica,

        I see! this was really helpful