How can I use Awk inside a Perl script?

Question

I'm having trouble using the following code inside my Perl script, any advise is really appreciated, how to correct the syntax?

# If I execute in bash, it's working just fine

bash$ whois google.com | egrep "\w+([._-]\w)*@\w+([._-]\w)*\.\w{2,4}" |awk ' {for (i=1;i<=NF;i++) {if ( $i ~ /[[:alpha:]]@[[:alpha:]]/ )  { print $i}}}'|head -n1

[email protected]

#-----------------------------------

#but this doesn't work 

bash$ ./email.pl google.com
awk:  {for (i=1;i<=NF;i++) {if (  ~ /[[:alpha:]]@[[:alpha:]]/ )  { print }}}
awk:                              ^ syntax error

# Here is my script
bash$ cat email.pl 
####\#!/usr/bin/perl         


$input = lc shift @ARGV;

$host = $input;

my $email = `whois $host | egrep "\w+([._-]\w)*@\w+([._-]\w)*\.\w{2,4}" |awk ' {for (i=1;i<=NF;i++) {if ( $i ~ /[[:alpha:]]@[[:alpha:]]/ )  { print $i}}}'|head -1`;
print my $email;

bash$

Answer 1

use a module such as Net::Whois if you want to code in Perl. Search CPAN for more such modules dealing with networking. If you want to use just Perl without a module, you can try this (note you don't have to use egrep/awk anymore, since Perl has its own grepping and string manipulation facilities )

   open(WHOIS, "whois google.com |")    || die "can't fork whois: $!";
   while (<WHOIS>) {

       print "--> $_\n";  # do something to with regex to get your email address
   }            
   close(WHOISE)                      || die "can't close whois: $!";

Answer 2

The easiest (though not the smoothest) way to use awk inside Perl is a2p.

echo 'your awk script' | a2p

Answer 3

Text inside backticks (or most* uses of the qx operator) is interpolated the same way as text inside double quoted strings, so it is easy to see what command you are sending to the shell:

$host = "google.com";
my $email_command = qq!whois $host | egrep "\w+([._-]\w)*@\w+([._-]\w)*\.\w{2,4}" |awk ' {for (i=1;i<=NF;i++) {if ( $i ~ /[[:alpha:]]@[[:alpha:]]/ )  { print $i}}}'|head -1!;
print $email_command;

---

Result (newlines included for readability):

whois google.com | egrep "w+([._-]w)*@w+([._-]w)*.w{2,4}" |
awk ' {for (i=1;i<=NF;i++) {if (  ~ /[[:alpha:]]@[[:alpha:]]/ )  { print }}}'|
head -1

To salvage your backtick'ed expression, at a minimum you need to double escape the metacharacters in the regular expression and escape the sigil in the awk variable $i.

* - the exception is when the qx token is the single quote char: $env_xxx = qx'echo $xxx'

Answer 4

As mentioned by others, backticks interpolate, so its tripping on the $'s. You could escape them all, or you could use single quotes like so:

open my $pipe, "-|", q[whois google.com | egrep ... |head -n1];
my $result = join "", <$pipe>;
close $pipe;

This takes advantage of open's ability to open a pipe. The -| indicates the filehandle $pipe should be attached to the output of the command. The chief advantage here is you can choose your quoting type, q[] is equivalent to single-quotes and not interpolated so you don't have to escape things.

But oh god, pasting an awk script into Perl is kind of silly and brittle. Either use a module like Net::Whois, or do the whois scraping in Perl, possibly taking advantage of things like Email::Find, or just write it as a bash script. Your Perl script isn't doing much in Perl as it stands.