dataDrone has asked for the wisdom of the Perl Monks concerning the following question:

Hey Folks... Any one else have trouble with using the Headers method of extracting a table from a page? I've narrowed it down to this: If two columns have similar headers TableExtract chokes. My HTML is a follows.
DateDate of Incident Unit cost
etc.... and when i make an array
@heads = ("Date","Date of Incident","Unit Cost"); $headers = \@heads ; $te = HTML::TableExtract->new(headers => $headers) ; $te->parse($page);
I get Bubkiss. If i change the HTML and @heads array to "Date", "Rate of Incident" , and "Unit Cost" the Regex pattern match works fine. Also if i change to Date_of_Incident is works. Unfortunately I don't have control of the generation of the HTML. I have gotten around the problem by using brute force and Liberal doses of Ignorance as follows.
@heads = ("myDate", etc... $page =~ s/Date/myDate/gs ; $te->parse($page);
Any suggestions? Thanks, The Drone.

Edit: footpad, ~Tue Nov 20 15:21:52 2001 (UTC)

Replies are listed 'Best First'.
Re: HTML::TableExtract revisited
by mkmcconn (Chaplain) on Nov 20, 2001 at 05:55 UTC

    That's a very curious problem, dataDrone. Apparently, the module will only succeed if it partly fails. For example, just switching the order of your headers makes the silly thing work.:

    #!/usr/bin/perl -w use strict; my $row; use HTML::TableExtract; $/ = ""; my $string = <DATA>; my $headers = ['Date of Incident','Date','Unit cost']; my $te = HTML::TableExtract->new(headers => $headers) ; $te->parse($string); # Examine all matching tables foreach my $ts ($te->table_states) { print "Table (", join(',', $ts->coords), "):\n"; foreach $row ($ts->rows) { print join(',', @$row), "\n"; } } # use Data::Dumper ; # print Dumper($te); __DATA__ <TABLE> <tr> <th>Date</th> <th>Date of Incident</th> <th>Unit cost</th></tr> <tr> <td>11-19-2001</td> <td>11-14-2001</td> <td>.01</td> </tr> <tr> <td>11-18-2001</td> <td>11-14-2001</td> <td>.01</td> </tr> </TABLE>

    The explanation might be found if you uncomment those Data::Dumper lines above, and examine the output, where the pattern to be tested is printed out. A match is tested against this pattern:
    'hpat' => '((Date of Incident)|(Date)|(Unit cost))'
    But I gave it no further thought beyond this. Good luck.

      Ah Yes.... having the hpat=>((Date of Incident)|(Date)|(Unit cost)) changes the way the pattern gets matched to the $self->htxt string. I think (very much not sure) that this is a Left associative regex problem. any Regex Gurus care to pitch in?