Why don't you provide an updated snippet of code for us to play with, and some sample HTML that results in the error. Just wrap the HTML in code tags. It's easier to debug an error that we can easily reproduce.

Also, you may want to use the 'debug' method from HTML::TableExtract to inspect the assertions your code makes about the state of affairs immediately before the call to 'rows'.

The code I'm sending the command prompt is the following: <\p>

 
#!/usr/bin/perl
use 5.014;    # so push/pop/etc work on scalars (experimental)
use strict;
use warnings;

use LWP::Simple 'get';
use HTML::TableExtract;


my $file = 'C:\Payout Policy Paper\Data\urllist.csv';
open (FH, "< $file") or die "Can't open $file for read: $!";
my @lines = <FH>;
close FH or die "Cannot close $file: $!"; 
print @lines;

foreach my $line (@lines) {


my $te = HTML::TableExtract->new(
    headers => [ 'Purchased','Average','Publicly','May'],
    slice_columns => 0,keep_html => 0,br_translate => 0
);

$te->parse($line);

my $table = $te->first_table_found;

my $file = "testout.csv";
open (F,">", $file);

for my $row ($table->rows) {
    print F join('^', @$row), "\n";
}

close (F);
}
[download]

If I'm correctly understanding your suggestion you can find an example of the html I am analyzing at:

 http://www.sec.gov/Archives/edgar/data/826083/000082608312000011/dellq1fy1310q.htm

Christoforo is correct (below): You're never actually fetching the URL. LWP::Simple's "get" is never invoked. Even if you're reading the table from a file, you probably shouldn't be parsing it "line by line". Slurp the file into a scalar and parse the whole thing.




Dave