Set difference with XPath 1.0 - how do I get .//table without .//table//table?
I'm trying to find all the tables below my current node without also including the nested tables. In othe开发者_运维技巧r words, if I have this, i want to find "yes" and not "no":
<table> <!-- outer table - no -->
<tr><td>
<div> <!-- *** context node *** -->
<table> <!-- yes -->
<tr><td>
<table> ... </table> <!-- no -->
</td></tr>
</table>
<table> <!-- yes -->
<tr><td>
<table> ... </table> <!-- no -->
</td></tr>
</table>
</div>
</td></tr>
</table>
Is there any easy way to do this in XPath 1.0? (In 2.0, it'd be .//table except .//table//table
, but I don't have a 2.0 as an option.)
EDIT: please, the answers so far are not respecting the idea of current context node. I don't know how far down the first layer of table might be (and it might differ), and I also don't know if I might be inside another table (or two or three).
Literally, I want what .//table except .//table//table
in XPath 2.0 would be, but I have only XPath 1.
I think you want child::table aka table
#!/usr/bin/perl --
use strict;
use warnings;
use HTML::TreeBuilder;
{
my $tree = HTML::TreeBuilder->new();
$tree->parse(<<'__HTML__');
<table> <!-- outer table - no -->
<tr><td>
<div> <!-- *** context node *** -->
<table> <!-- yes -->
<tr><td>
<table> ... </table> <!-- no -->
</td></tr>
</table>
<table> <!-- yes -->
<tr><td>
<table> ... </table> <!-- no -->
</td></tr>
</table>
</div>
</td></tr>
</table>
__HTML__
sub HTML::Element::addressx {
return join(
'/',
'/', # // ROOT
reverse( # so it starts at the top
map {
my $n = $_->pindex() || '0';
my $t = $_->tag;
$t . '['. $n .']'
} # so that root's undef -> '0'
$_[0], # self and...
$_[0]->lineage
)
);
} ## end sub HTML::Element::addressx
for my $td ( $tree->look_down( _tag => qr/div|table/i ) ) {
print $td->addressx, "\n";
}
$tree->delete;
undef $tree;
}
__END__
//html[0]/body[1]/table[0]
//html[0]/body[1]/table[0]/tr[0]/td[0]/div[0]
//html[0]/body[1]/table[0]/tr[0]/td[0]/div[0]/table[0]
//html[0]/body[1]/table[0]/tr[0]/td[0]/div[0]/table[0]/tr[0]/td[0]/table[0]
//html[0]/body[1]/table[0]/tr[0]/td[0]/div[0]/table[1]
//html[0]/body[1]/table[0]/tr[0]/td[0]/div[0]/table[1]/tr[0]/td[0]/table[0]
and second part
#!/usr/bin/perl --
use strict;
use warnings;
use HTML::TreeBuilder::XPath;
my $tree = HTML::TreeBuilder::XPath->new;
$tree->parse_content(<<'__HTML__');
<table> <!-- outer table - no -->
<tr><td>
<div> <!-- *** context node *** -->
<table> <!-- yes -->
<tr><td>
<table> ... </table> <!-- no -->
</td></tr>
</table>
<table> <!-- yes -->
<tr><td>
<table> ... </table> <!-- no -->
</td></tr>
</table>
</div>
</td></tr>
</table>
__HTML__
#~ for my $result ($tree->findnodes(q{//html[0]/body[1]/table[0]/tr[0]/td[0]/div[0]})) {
for my $result ($tree->findnodes(q{/html/body/table/tr/td/div})) {
print $result->as_HTML,"\n\n";
for my $table( $result->findnodes(q{table}) ){ ## child::table
print "$table\n";
print $table->as_HTML,"\n\n\n";
}
}
__END__
<div><table><tr><td><table><tr><td> ... </td></tr></table></td></tr></table><table><tr><td><table><tr><td> ... </td></tr></table></td></tr></table></div>
HTML::Element=HASH(0xc6c964)
<table><tr><td><table><tr><td> ... </td></tr></table></td></tr></table>
HTML::Element=HASH(0xc6cbf4)
<table><tr><td><table><tr><td> ... </td></tr></table></td></tr></table>
Well, if I understand it, the content_list can solve:
my $table_one = $tree->findnodes('/html//table')->[1];
for ( $table_one->content_list ) {
last if $_->exists('table');
print $_->as_text;
}
:)
What about .//table[not(.//table)]
? Sorry for brevity, I'm on my phone.
I don't know how to get the context node to be evaluated in the nested predicates, but what you need is something like this:
descendant::table[not(ancestor::table[ancestor::div])]
only with the ability to reference the context node, instead of div
EDIT: If you set a variable for the context node,
<xsl:variable name="contextNode" select="." />
then you can reference it in the XPATH predicate:
descendant::table[not(ancestor::table[ancestor::*[generate-id(.)=generate-id($contextNode)]])]
After investigating it here and elsewhere, the answer seems to be "you can't, and that's why we have XPath 2.0". Oh well.
精彩评论