Help to extract lines from email file in Perl

[Paul D. Simpson]

Hi all,
Thanks in advance for the help.

I have a Perl script that is grabbing headers from email messages & logging them to a file... Unfortunately, my sender switched the mechanism he uses to send these emails & now there seems to be a newline on some of the data I am trying to grab (Subject is the main culprit). I'd like to make my script grab everything from "Subject:" up to right before "To:", and "To:" up to just before "Cc:", "Cc:" up to just before "Message-ID"

Can anyone help please?

My script contains:
#!/opt/perl/bin/perl -w

while ($line=) {
my @now = localtime();
my $timestamp = sprintf("%04d-%02d-%02d",$now[5]+1900, $now[4]+1, $now[3]);
my $OUTPUT = "/home/share/acks/order_acks_$timestamp.txt";
open (OUTFIL, ">>$OUTPUT") || die "Cannot open $OUTPUT file";
if ($line =~ /^To:/){
chomp $line;
print OUTFIL "$line \n";
}
if ($line =~ /^[Cc]c:/){
chomp $line;
print OUTFIL "$line \n";
}
if ($line =~ /^Subject:/){
chomp $line;
print OUTFIL "$line \n\n";
}
}
close OUTFIL;
exit;
(Yes, I'm new to Perl)


The file I'm grabbing from looks like:
( I added ###<---->###### tages to highlight what I need)

From Sales.Order.Ack.Exceptions@company.com Wed Jun 10 15:56:21 2009
Return-Path:
Received: from smtp02.company.com (chn-sv-exft02.microchip.com [99.99.99.18])
by newton.company.com (8.9.3 (PHNE_35484)/8.9.3) with ESMTP id PAA03316;
Wed, 10 Jun 2009 15:56:21 -0700 (MST)
Received: from CHN-VM-STRMSRV.company.com ([99.99.99.126]) by smtp02.company.com with Microsoft SMTPSVC(6.0.3790.1830);
Wed, 10 Jun 2009 15:56:20 -0700
Date: Wed, 10 Jun 2009 15:56:20 -0700 (PST)
From: Your Order Acknowledgement
Reply-To: Sales.Order.Ack.Exceptions@company.com
####<-- SUBJECT GRAB START HERE-->#####
Subject: =?ISO-8859-1?Q?Customer:_110557,_Customer_PO:_123456789,_Order:_132456(?=
=?ISO-8859-1?Q?Changed_lines_41)__06-10?=
####<-- SUBJECT GRAB END HERE-->#####
####<-- TO GRAB START HERE-->#####
To: customer456@yahoo.com
####<-- TO GRAB END HERE-->#####
####<-- CC GRAB START HERE-->#####
####<-- CC GRAB END HERE-->#####
Message-ID: <7081244674580691@company.com>
MIME-Version: 1.0
Content-Type: MULTIPART/MIXED; BOUNDARY="6978513-22148-1244674580=:708"
X-OriginalArrivalTime: 10 Jun 2009 22:56:20.0044 (UTC) FILETIME=[AB49F8C0:01C9EA1E]

--6978513-22148-1244674580=:708
Content-Type: TEXT/PLAIN; CHARSET=ISO-8859-1
Content-Transfer-Encoding: QUOTED-PRINTABLE

We thank you for your order. An Order Acknowledgement is attached.=0A=0AIf=
you need further information...

[James R. Ferguson · Accepted solution]

Hi Paul:

If the lines are truly multiple ones, we could do something like this:

# cat ./mymail
#!/usr/bin/perl
use strict;
use warnings;
my $line;
my @now = localtime();
my $timestamp
= sprintf( "%04d-%02d-%02d", $now[5] + 1900, $now[4] + 1, $now[3] );
my $OUTPUT = "home/share/acks/order_acks_$timestamp.txt";

#->open (OUTFIL, ">>$OUTPUT") || die "Cannot open $OUTPUT file";
{
local $/ = undef;

while ( $line = <> ) {
chomp $line;
if ( $line =~ /( ^To:.*? )( [Cc]c:|Message-ID )/xms ) {
print $1, "\n";
}
if ( $line =~ /( ^[Cc]c:.*? )( Message-ID )/xms ) {
print "$line \n";
}
if ( $line =~ /( ^Subject:.*? )( To: )/xms ) {
print $1, "\n";
}
}
}
close OUTFIL;
exit;

...

We slurp the whole file into one big record. Then we capture the pieces we want by matching the across any newline characters with '/s' regex modifier to allow '.' to match a newline.

Regards!

...JRF...

[Paul D. Simpson]

Thanks James,

It is definitely sending the correct output to the screen, but does not seem to be writing to the output file.

FYI: This is being used as an alias redirect for sendmail, such that incoming mail for address "ack" is piped to this script. Not sure if that makes a difference.

Paul

[James R. Ferguson]

Hi:

> It is definitely sending the correct output to the screen, but does not seem to be writing to the output file.

Yes, that's because for ease of testing I dropped your output filehandle leaving STDOUT as the default, selected one.

Instead of:

# print $1, "\n";

change to:

# print OUTFIL $1, "\n";

(or):

# print OUTFIL "$1\n";

Regards!

...JRF...

[Paul D. Simpson]

James,
Thanks... I should have caught that... I had the same problem when I first wrote the script.
Thanks again
-Paul

[Paul D. Simpson]

James nailed it