Extracting nested tables from HTML

Hi!

I have several very large HTML files from which I'd like to extract only
tables nested at the deepest level.  I thought this would be quite easy
by extracting  something like (<table.*?</table> ) where I'd alter the
'.*?' to test for and reject any new occurrences of a starting table
tag, but I can't seem to get it.  Any pointers?

I want to deal with the file at a text level until the tables are
extracted, after which I plan to use HTML::TableContentParser to extract
the needed content.

Thanks for your help.

Terry.

"Terry" <just@say.no> wrote in message
 news:og8at0tfs8veatosp6r9i1u1dmb09jqcbb@
4ax.com...
> Hi!
>
> I have several very large HTML files from which I'd like to extract
only
> tables nested at the deepest level.  I thought this would be quite
easy
> by extracting  something like (<table.*?</table> ) where I'd alter the
> '.*?' to test for and reject any new occurrences of a starting table
> tag, but I can't seem to get it.  Any pointers?
>
> I want to deal with the file at a text level until the tables are
> extracted, after which I plan to use HTML::TableContentParser to
extract
> the needed content.

Search the archives of this newsgroup for "HTML" and "Regular
Expression" to see why this is simply a bad idea.  Regular expressions
are not sufficient to parse HTML.  That is why CPAN contains so many
HTML parsers.   Visit http://search.cpan.org and search for HTML, then
try one that looks like it best suits your needs.

Paul Lalli

Terry <just@say.no> wrote:


> I have several very large HTML files


How large might "very large" be when you say it?

In my experience, only very poorly designed websites serve what I
would call "very large" HTML files...


> from which I'd like to extract only
> tables nested at the deepest level.


Is this deepest level an arbitrary depth, or is it always the
same depth for this type of HTML file?

The answer to this would have a large impact on the approaches
available.


> I thought this would be quite easy
> by extracting  something like (<table.*?</table> )


I guess you've never taken a Formal Methods class then.  :-)


> where I'd alter the
> '.*?' to test for and reject any new occurrences of a starting table
> tag, but I can't seem to get it.  Any pointers?


Yes, but you have surely already seen them since they are
Questions that are Asked Frequently:

perldoc -q nest

How do I find matching/nesting anything?

perldoc -q HTML

How do I remove HTML from a string?


> I want to deal with the file at a text level until the tables are
> extracted, after which I plan to use HTML::TableContentParser to extract
> the needed content.


I would use HTML::TableExtract.

It gives you the depth via the coords() method.


--
Tad McClellan                          SGML consulting
tadmc@augustmail.com                   Perl programming
Fort Worth, Texas