Dom4j xpath problem

Question

Hello all
I have several methods in which I noticed strange behavior
Here is the example:
I create xpath as object because of performance, cause i use it on thousands of files
XPath pathSpanPodzakonska = DocumentHelper.createXPath("//span[@class='palnk']");
I found h4 element via xpath with relative path to h4 element...
That works fine so far and founds h4 element that I need to.
final Element elementH4= (Element) document.selectSingleNode("//h4[@id='someId']");
Now i need to found some element inside that h4 tag
but what I get is first element in document, outside h4 element !!!
final Node nodeSpanPodzakonska = pathSpanPodzakonska.selectSingleNode(elementH4);
That works same even if i write
final Node nodeSpanPodzakonska = elementH4.selectSingleNode("//span[@class='palnk']");
If I write code with absolute xpath that works fine when the span is directly inside of h4 but this isn't always case and cant depend on that.
And more important have some other similar methods and want to clarify what happens in this xpath implementations and how can I use it in the best way.
final Node nodeSpanPodzakonska = elementH4.selectSingleNode("span[@class='palnk']");
Danilo Cubrovic

Answer 1

Sorry. I will have to think about the XPath question in more detail.
As to the side by side comparison: I meant of - open source HTML parsers
http://java-source.net/open-source/html-parsers
and also of XML frameworks that support XPath (such as DOM4j vs XOM).
I am going to try HTMLParser with XOM, as I have, in informal speed testing, seen HTMLParser to be quite fast compared to others.
The reason I would like to have speed is that I am working on a Web automation toolkit:
www.mkosh.com
and it looks like I will be re-parsing the entire tree after many Javascript commands.
I would like to make this as speedy as possible. HTMLParser seems to win hands down, although if it does not work with real world HTML, that could be an issue.
I will have to test.
Thank you so much
Misha