use WWW::Mechanize;
use HTML::TableExtract;

$url="http://www.dhl-usa.com/TransitTimes/USTTimeStart.asp?nav=Transit
+Times";
$service = "Service";
$arrival = "Arrival Date and Time";
$transit = "Days in Transit*";
my $mech = WWW::Mechanize->new();
$mech->get($url);

$frmSvcCalc = "frmSvcCalc";
$mech->form_name($frmSvcCalc);
$mech->set_fields(
                  txtOrgZip => "53213",
                  txtDestZip => "60056",
                  );
$mech->field( "hdnAction", "Calculate" );

$mech->submit($frmSvcCalc);
my $results2 = $mech->content;

$te = HTML::TableExtract->new( headers => ['Service','Arrival Date and
+ Time','Days in Transit*'] );
$te->parse($results2);

foreach $ts ($te->tables) {
    foreach $row (@$ts) {
        print join(',', @$row), "\n";
    }
}
[download]

Update I was using version 1.08 and looking at docs on search.cpan.org for 2.06, take with a grain of salt :)

Thinking about it, I should give you more info on what I did, so you can do it yourself in the future :)

Personally I was just using print debugging. I started by getting

Can't call method "coords" on unblessed reference at test.pl line 32.
[download]

I printed out the value of $ts, which was an array, but was not blessed to a package. Since it had no package associated with it (when printing, HASH(0x864497c) is not blessed, HTML::TableExtract=HASH(0x864497c) is) that means you cannot call methods on it. So then I did print "@$ts\n"; and saw it was full of array refs. At this point you could loop through the array and print the results or do a print "@{$ts->[0]}\n";. Of course... the much better way to do this is use Data::Dumper and do a print Dumper($ts),"\n"; which would give you

$VAR1 = [
          [
            'DHL Next Day 10:30 am
 (Letter – 150 Pounds)',
            'Tuesday, Jan 10, 2006  By 10:30 A.M.',
            '1'
          ],
          [
            'DHL Next Day 12:00 pm
 (Letter – 150 Pounds)',
            'Tuesday, Jan 10, 2006  By Noon',
            '1'
          ],
          [
            'DHL Next Day 3:00 pm
 (Letter – 150 Pounds)',
            'Tuesday, Jan 10, 2006  By 3:00 P.M.',
            '1'
          ],
          [
            'DHL 2nd Day Service
 (Letter – 150 Pounds)',
            'Wednesday, Jan 11, 2006  By 5:00 P.M.',
            '2'
          ],
          [
            'DHL Ground Service
 (Letter – 150 Pounds)',
            'Tuesday, Jan 10, 2006  By end of day',
            '1'
          ]
        ];
[download]

Which shows the whole structure obviously.... quick and dirty debugging can often help you find out what is up, or you could always actually use the debugger, which I am usually too lazy to do :)

Hopefully this helps even more than the first message.

Any questions?

                - Ant
                - Some of my best work - (1 2 3)

my $te = HTML::TableExtract->new( headers => [qw($service $arrival $tr
+ansit)] );
$te->parse($html_string);
   foreach my $row ($te->rows) {
    foreach my $cell (@$row) {
      print $cell;
    }
}
[download]

I note that in your example code, you're not setting $html_string beforehand. That could be a paste-o, but make sure you're using strict to catch that kind of thing.

Having said all that -- if it's still not working, which version of HTML::TableExtract are you using?

Thanks,
Matt (author of said module)

Cool module :)

I looked at my version and I have 1.08, for simplicity I had just gone with the debian version, I guess its a bit behind.

I wonder if he did the same as me, has an older version but was looking at the docs on search.cpan.org :)

                - Ant
                - Some of my best work - (1 2 3)

With 1.08, try this:

my $te = HTML::TableExtract->new( headers => [qw($service $arrival $tr
+ansit)] );
$te->parse($html_string);
my $ts = $te->first_table_state_found;
foreach my $row ($ts->rows) {
  foreach my $cell (@$row) {
    print $cell;
  }
}
[download]

If it's actually capturing tables, that should work. (you can monitor, in detail, whether tables are being captured by setting the 'debug' parameter to varying levels (1..7) in the new() constructor).

Let me know what happens,
Matt

And yes, part of this is my fault. As I phased out the array-only functionality (i.e. the tables() method) I eventually aliased that to the table_states() method and it has caused confusion. I should have waited longer before ressurecting the tables() method, if at all. (in my defense, it is what I would have called the table_states() method to begin with, but in a very very early version of the module I used tables() to return the arrays. Ah well.)

Cheers,
Matt

use strict;
use LWP::Simple;
use HTML::TableExtract;


my $page = get("http://www.dhl-usa.com/TransitTimes/USTTimeRslts.asp?n
+av=TransitTimes&oz=53213&oc=1&oh=ORD&dz=60056&dc=1&dt=1/9/2006&tt=1&h
+y=&zn=2&am=Y");

my $te = HTML::TableExtract->new( depth => 3, count => 3);
$te->parse($page);

foreach my $ts ($te->table_states) {
   foreach my $row ($ts->rows) {
      print join(',', @$row), "\n";
   }
 }

__OUTPUT__
Service,Arrival Date and Time,Days in Transit*
DHL Next Day 10:30 am
 (Letter ű 150 Pounds),Tuesday,áJaná10,á2006 áBy 10:30 A.M.,1
DHL Next Day 12:00 pm
 (Letter ű 150 Pounds),Tuesday,áJaná10,á2006 áBy Noon,1
DHL Next Day 3:00 pm
 (Letter ű 150 Pounds),Tuesday,áJaná10,á2006 áBy 3:00 P.M.,1
DHL 2nd Day Service
 (Letter ű 150 Pounds),Wednesday,áJaná11,á2006 áBy 5:00 P.M.,2
DHL Ground Service
 (Letter ű 150 Pounds),Tuesday,áJaná10,á2006 áBy end of day,1
[download]