Howto insert awk command in perl script?

Question

I want to add this awk command on my script but keep getting error. I have put inside " " but still getting error.

system("awk -F"\t" '{ for ( i=1; i<=2; i++ ) { printf "%s\t", $i } printf "\n"; }' myfile file2"};

the errors are

String found where operator expected at host_parse line 21, near "t" '{ for ( i=1; i<=2; i++ ) { printf ""

Unquoted string "a" may clash with future reserved word at myfile line 58.

Unquoted string "a" may clash with future reserved word at myfile line 58.

syntax error at myfile line 21, near "" awk -F"\\"

Thanks.

Answer 1

One of the trickiest parts about using the system command is using quotes in a way that will get the correct command passed to the operating system. Perl's q// construction can be very helpful for this:

# treat everything between the @...@ as uninterpolated string
system( q@awk -F"\t" '{ for ( i=1; i<=2; i++ ) { printf "%s\t", $i } 
          printf "\n"; }' myfile file2@ );

Answer 2

To answer your immediate question, you're tripping over the default behavior of Perl's system operator. Usually, it's a great convenience for the shell to parse the command, but sometimes, as as you've seen, having multiple levels of encoding is a pain—or even a security vulnerability .

You can bypass the shell's quoting entirely with the system LIST and exec LIST forms. In your case, change your code to

#! /usr/bin/env perl

use strict;
use warnings;

my @cmd = (
  "awk",
  "-F", "\t",
  '{ for ( i=1; i<=2; i++ ) {
       printf "%s\t", $i
     }
     printf "\n";
   }',
   "myfile", "file2",
);

system(@cmd) == 0 or warn "$0: awk exited " . ($? >> 8);

You don't have to use the temporary array, but I don't like the resulting code with a multi-line command and a check for success.

Given myfile containing

1  2   3   4
foo bar baz
oui oui monsieur

and file2 with

a   b   c
d   e   f   g

(where the separators in both cases are TAB characters), then the output is

1  2   
foo bar 
oui oui 
a   b   
d   e

They're invisible, but each line of output above has a trailing TAB.

Doing the same in Perl is straightforward. For example,

sub print_first_two_columns {
  foreach my $path (@_) {
    open my $fh, "<", $path or die "$0: open $path: $!";

    while (<$fh>) {
      chomp;
      my(@cols) = (split /\t/)[0 .. 1];
      print join("\t", @cols), "\n";
    }

    close $fh;
  }
}

The part that may not be obvious is taking a slice of the values returned from split , but what's happening is simple in concept. A slice allows you to grab data at multiple indices (0 and 1 in this case, ie , the first and second columns). The range-operator expression 0 .. 1 evaluates to the list 0 and 1. If you decide later you want the first four columns, you'd change it to 0 .. 3 .

Call the sub above as in

print_first_two_columns "myfile", "file2";

Note that the code isn't exactly equivalent: it doesn't preserve the trailing TAB characters.

From the command line, it's even simpler:

$ perl -lane '$,="\t"; print @F[0,1]' myfile file2 
1   2
foo bar
oui oui
a   b
d   e

Answer 3

You don't need the shell to interpret any redirection (or other shell facilities), so it would be better to pass a list of arguments to system()

system 'awk', '-F', "\t", 
   '{for (i=1; i<=2; i++) {printf "%s\t", $i}; print ""}',
   'myfile', 'file2';