From joshua.mcadams at gmail.com  Thu Jun  1 08:13:59 2006
From: joshua.mcadams at gmail.com (Joshua McAdams)
Date: Thu, 1 Jun 2006 11:13:59 -0400
Subject: [Classiccity-pm] YAPC::NA
Message-ID: <49d805d70606010813r41631e5arddf7f1e4a4cae8e0@mail.gmail.com>

Hi there fellow Perl Mongers.  I'm writing to remind you all that
YAPC::NA is only a few weeks away.  The conference will be held in
Chicago June 26th through 28th and will feature four simultaneous
sessions of Perl talks for three days in addition to a job fair,
banquet, and auction.  After the conference Damian Conway, Randal
Schwartz, and brian d foy will be sticking around and conducting
professional training classes and extremely reduced prices.

This email is a little spammy (sorry about that), but I just wanted to
remind you all about the conference and also ask for your help in
promoting it so that we can fill up the few spots that are remaining.

For more information check out http://www.yapcchicago.org.

We invite you to put up posters:

 http://yapcchicago.org/yapc_poster.pdf
 http://yapcchicago.org/yapc_poster_white.pdf

Or maybe a web banner:

 http://www.yapcchicago.org/yapc_banner_wide.jpg
 http://www.yapcchicago.org/yapc_banner_narrow.jpg

Thank you for your help in making YAPC a success once again,

Josh McAdams

From lindadl at uga.edu  Mon Jun  5 11:02:16 2006
From: lindadl at uga.edu (Linda Dubes Law)
Date: Mon, 5 Jun 2006 14:02:16 -0400
Subject: [Classiccity-pm] Greatest number
Message-ID: <20060605140216890.00000002976@lindadl>

I am trying to figure out an efficient way to get the greatest number in a file.  The file is rather large.  Should I do something like:
*********************
while (<DPCARDS>) {
	chomp;
	($cardnumber) = (split / /) [1];
	if ($cardnumber gt $largest) {
		$largest = $cardnumber
*********************

Or should I read everything in an array and then sort the array 

Or should I go to the system, sort the file and then is there some way to read in only the last record of a file?  Or perhaps sort descending and then read in only the first record?

As always, your input is appreciated!

Thanks, 

Linda Law


From joshua.mcadams at gmail.com  Mon Jun  5 11:12:19 2006
From: joshua.mcadams at gmail.com (Joshua McAdams)
Date: Mon, 5 Jun 2006 14:12:19 -0400
Subject: [Classiccity-pm] Greatest number
In-Reply-To: <20060605140216890.00000002976@lindadl>
References: <20060605140216890.00000002976@lindadl>
Message-ID: <49d805d70606051112h15a68e37h4b78f28ab1a4e864@mail.gmail.com>

> Or should I read everything in an array and then sort the array

I would think that if file size might be an issue, then you'd probably
want to stick with what you are doing now and not load the entire
thing into memory.

Just to be sure, you know that you're doing a string compare on the
numbers instead of a numeric compare?  Probably not too big of a deal
and maybe even what you intended... just thought I'd call it out.

From pkeck at uga.edu  Mon Jun  5 11:22:08 2006
From: pkeck at uga.edu (Paul Keck)
Date: Mon, 5 Jun 2006 14:22:08 -0400
Subject: [Classiccity-pm] Greatest number
In-Reply-To: <20060605140216890.00000002976@lindadl>
References: <20060605140216890.00000002976@lindadl>
Message-ID: <20060605182208.GI4200@uga.edu>

On Mon, Jun 05, 2006 at 02:02:16PM -0400, Linda Dubes Law wrote:
> Or should I go to the system, sort the file and then is there some way to
> read in only the last record of a file?  Or perhaps sort descending and
> then read in only the first record?

tail -1
head -1

will give you the head and tail line.  I'd try running it like

system("cat filename | sort -n +3 -4 |tail -1")

and adjust the sort until it sorted on the proper field.  See how long that
takes compared to using perl.


If you need to do it in perl alone, then the code you put in looked pretty
good to me.  No need to store the lines in an array if that's all you want
out of there.  If you're looking at making multiple passes through it, I
suggest sucking it into a hash.

-- 
Paul Keck       pkeck at uga.edu         http://www.arches.uga.edu/~pkeck
University of Georgia                 http://www.uga.edu/ucns/telecom
EITS Network Engineering              mailto:pkeck at ediacara.org
    --Opinions mine.--                Go fighting anomalocaridids!!!

From dwestbrook at tungstenlearning.com  Mon Jun  5 12:22:48 2006
From: dwestbrook at tungstenlearning.com (David Westbrook)
Date: Mon, 05 Jun 2006 15:22:48 -0400
Subject: [Classiccity-pm] Greatest number
In-Reply-To: <20060605182208.GI4200@uga.edu>
References: <20060605140216890.00000002976@lindadl>
	<20060605182208.GI4200@uga.edu>
Message-ID: <44848488.9060003@tungstenlearning.com>

Paul Keck wrote:

> system("cat filename | sort -n +3 -4 |tail -1")

That's a "useless use of cat" :)    
http://laku19.adsl.netsonic.fi/%7Eera/unix/award.html

What do the "+3 -4" to sort do?

A blind guess might be that "head -1" is faster than "tail -1" on a pipe 
so it doesn't have to read to the end (unless it has someway to seek to 
the end), so using sort's "-r" option might increase it a little.
Also need backticks instead of system() to get the output ..
And i'll throw in an awk since OP wanted 2nd column (could use cut if 
the actual delim was known):

my $max_from_shellout = `awk '{print $2}' | sort -rn $filename | head -1`


Cool -- after a "man awk" (i've only ever used it in a "cut" fashion), 
it can sort as well:
  awk '{ print $1 | "sort" }' /tmp/n | tail -1


A pure-perl (and possibly overkill, but maybe not -- no idea of context 
here, but maybe this would be a good way for OP to access/update the 
file in question) way would be to use Class::DBI and DBD::CSV to set up 
a db object wrapper class for this "table" (file), and just use the 
maximum_value_of() method.
package My::Sample;
use base 'Class::DBI';
__PACKAGE__->connection("DBI:CSV:f_dir=data/");
__PACKAGE__->table('player_changes');
__PACKAGE__->columns( All => qw/ col1 col2 foo bar / );
############
in code:   my $max = My::Sample->maximum_value_of('col2');
(maybe not the most effiicient either, but fun, right? ;) )

--david