while (my $line = <INPUT>) {
    my ($time,$email,$name,@addr) = split(/\|/,$line);
    # use $email for cool stuff here
    # or add to an array for mailing.
}
[download]

I really, sincerely hope that this is not real data. I would be terribly upset to see my name, e-mail and full address posted on some random web page.

Please respond or update your post to acknowledge that this data is fabricated, or I must recommend to the monks that this post be sanitized.

Update: Thanks to Masem for cleaning the code - I hadn't realized that a root SoPW node couldn't be updated (thanks to ChemBoy for pointing that out). I work in the healthcare field and am forced to think constantly about protecting patient data, so forgive me if I over-reacted.

Sincerely,
{NULE}
--
http://www.nule.org

Personally, I don't think you over-reacted. We should all take great to protect personal data as much as possible. I think this is a good reminder for us all.

I have a link in a bookmark file lying around somewhere, a Usenet post, recorded Deja/Google's archive. Some poor fool posted his credit card number on a newsgroup.

Fortunately we can rectify such mistakes here on this site. On Usenet, you're out of luck.

--

g r i n d e r

It is not real data. I changed it to protect the innocent (or something like that).

Thanks for looking out for dumb mistakes. I know I've made plenty. I will make sure in the future that I include verbage to let you know that it is not real.

peppiv

Greetings peppiv,

If all you want to do is take data from one file and dump it into another, just setup your filehandles and perform the following:

The regex on the grep isn't all that great. merlyn has a few that work much better for finding all sorts of valid email addresses. However, for your example data, this works just fine.

So here's how it works (starting right, moving left):

1. Grab the file data and use map to split every field of every line into an element of an array.
2. Then we're going to grep that array for all elements that look like they might be email addresses.
3. Then we're going to join that array of maybe-valid email addresses into a single string with line-breaks.
4. Finally, print that to whatever output file you setup.

-gryphon
code('Perl') || die;

Please forgive me if I've misunderstood the question (and subsequent replies), but it sounds like searching for email addresses in the line isn't necessary -- that the email addresses are always the second element of the | separated line.

Here's a way to extract only the second element and dump it (one per line) to another file.

But if I may ask, why dump the email addresses into another file if you'll just need to re-open the new file to grab the email addresses for further processing? You can use this to grab the emails from your INFILE, then loop through them to perform your actual mailing:</P?

foreach my $email ( map { ( split( /\|/ ) )[1] } <INFILE> ){

    # Do something with $email

}
[download]

As you can see from the benchmark results below, the foreach loop isn't that much slower than throwing the data into another file, which will then need to be re-opened, re-read, re-closed.

Benchmark: timing 250000 iterations of foreachmapsplit, greprxmapsplit
+, mapsplit...

foreachmapsplit:  4 wallclock secs ( 3.46 usr +  0.00 sys =  3.46 CPU)
+ @ 72254.34/s (n=250000)
greprxmapsplit:  4 wallclock secs ( 3.24 usr +  0.00 sys =  3.24 CPU) 
+@ 77160.49/s (n=250000)
  mapsplit:  4 wallclock secs ( 3.24 usr +  0.00 sys =  3.24 CPU) @ 77
+160.49/s (n=250000)
[download]

The greprxmapsplit code is gryphon's. I included it because the grep-regex/map/split consistently matched the plain map/split. One day I'll understand why :)

# this is a lazy regexp - you can refine it better, I'm sure.
# $line is the line of text from your example
@emails = grep ( /.*\@.*/, split( /\|/, $line ) );
[download]

Jason

update - before the sanitization of the data above, it was all in one line. No, really! ;)