Reputation: 26272
XML:
<body>
<h2><font style="font-weight: bold">Baz</font></h2>
<p><img title="image" /></p>
<p>Baz 0 with an <a href="http://">anchor</a> element.</p>
<p>Baz 1 with an <a href="http://">anchor</a> element.</p>
<hr />
<h2><font style="font-weight: bold">People</font></h2>
<ul>
<li>People 0 with <a href="http://" >an anchor</a> element.</li>
<li>People 1 with an <a href="http://" >an anchor</a> element.</li>
</ul>
<hr/>
<h2><font style="font-weight: bold">Sales</font></h2>
<ul>
<li>List item 2 with an <a href="http://" >an anchor</a> element.</li>
<li>List item 3 with an <a href="http://" >an anchor</a> element.</li>
<li>List item 4 without an anchor element.</li>
</ul>
<hr />
<h2><font style="font-weight: bold">Sales</font></h2>
<p><img title="image" /></p>
<p>sales 0 with an <a href="http://">anchor</a> element.</p>
<p>sales 1 with an <a href="http://">anchor</a> element.</p>
<hr />
<h2><font style="font-weight: bold">Foo</font></h2>
<ul>
<li>Foo 0 with <a href="http://" >an anchor</a> element.</li>
<li>Foo 1 with an <a href="http://" >an anchor</a> element.</li>
</ul>
<hr />
<h2><font style="font-weight: bold">Bar</font></h2>
<p><img title="image" /></p>
<p>bar 0 with an <a href="http://">anchor</a> element.</p>
<p>bar 1 with an <a href="http://">anchor</a> element.</p>
<hr />
</body>
This xpath: //p[a and preceding-sibling::h2[font[text()='Sales']][1] and following-sibling::hr[1]]
Returns:
<p>sales 0 with an <a href="http://">anchor</a> element.</p>
<p>sales 1 with an <a href="http://">anchor</a> element.</p>
<p>bar 0 with an <a href="http://">anchor</a> element.</p>
<p>bar 1 with an <a href="http://">anchor</a> element.</p>
Desired p
:
<p>sales 0 with an <a href="http://">anchor</a> element.</p>
<p>sales 1 with an <a href="http://">anchor</a> element.</p>
Desired li
:
<li>List item 2 with an <a href="http://" >an anchor</a> element.</li>
<li>List item 3 with an <a href="http://" >an anchor</a> element.</li>
What am I missing?
How would I alter the xpath to include li/[a]
in the same manner that I'm including p/[a]
? The preceding/following-sibling
wouldn't work with li
.
Upvotes: 3
Views: 3029
Reputation: 52888
You should just need to specify that it's the first preceding-sibling h2:
preceding-sibling::h2[1]
updated xpath (I also simplified the test of Sales
):
//p[a and preceding-sibling::h2[1][.='Sales'] and following-sibling::hr]
Also, if you need to be certain that the first following-sibling that isn't p
is hr
, you can try this...
//p[a and preceding-sibling::h2[1][.='Sales'] and following-sibling::*[not(self::p)][1][self::hr]]
If you're trying to also select li
in addition to p
, you could update the xpath to use preceding::
and following::
, but you'd have to account for any elements that might appear as children of p
like a
, span
, etc...
//*[self::p or self::li][a and preceding::h2[1][.='Sales'] and following::*[not(self::p) and not(self::li) and not(self::a)][1][self::hr]]
This would select the following from your sample XML...
<li>List item 2 with an <a href="http://" >an anchor</a> element.</li>
<li>List item 3 with an <a href="http://" >an anchor</a> element.</li>
<p>sales 0 with an <a href="http://">anchor</a> element.</p>
<p>sales 1 with an <a href="http://">anchor</a> element.</p>
However, I'd recommend a second xpath to target li
specifically...
//li[a and preceding::h2[1][.='Sales'] and ../following-sibling::*[1][self::hr]]
Upvotes: 3