4
<h2>Headline 1</h2>
<p>some text</p>
<p>some more text</p>
<ul>
<li>list item 1</li>
<li>list item 2</li>
</ul>
<p>more text</p>
<h2>Headline 2</h2>

I have the above in a webpage and I want to be able to target all elements following the first h2 that contains the text 'Headline 1' up to but NOT including the element h2 that contains the text 'Headline 2'.

I can target the elements like this:

//*[count(preceding-sibling::hr)=1]

but this is not specific to the text contained and so if the page ever changed then the xpath could be pointing to something totally different.

What I would like in sudo code terms is this:

give me all the elements between the header 'Headline 1' and the header 'Headline 2' including 'Headline 1'

Is this at all possible?

kjhughes
  • 106,133
  • 27
  • 181
  • 240
Jason Hegarty
  • 41
  • 3

3 Answers3

3

This XPath,

//*[    preceding-sibling::h2[. = 'Headline 1'] 
    and following-sibling::h2[. = 'Headline 2']]

will select all elements between h2s with string values of 'Headline 1' and 'Headline 2':

<p>some text</p>
<p>some more text</p>
<ul>
<li>list item 1</li>
<li>list item 2</li>
</ul>
<p>more text</p>

Andersson points out in the comments that OP wants the first h2 included in the selection.

Andersson's initial thought would work:

//h2[. = 'Headline 1'] |
//*[    preceding-sibling::h2[. = 'Headline 1'] 
    and following-sibling::h2[. = 'Headline 2']]

Here's another way:

//*[self::h2[. = 'Headline 1']
    or (    preceding-sibling::h2[. = 'Headline 1'] 
        and following-sibling::h2[. = 'Headline 2']]

Or, probably the ideal way:

//h2[. = 'Headline 2']
    /preceding-sibling::*[not(following-sibling::h2[. = 'Heading 1'])]

because it avoids having to specify 'Heading 1' twice.

kjhughes
  • 106,133
  • 27
  • 181
  • 240
0

What if try something like 

//*[text()="Headline 1"]/following-sibling::*
Cristian Guillen
  • 21
  • 4
0
//h2[contains(.,'Headline 1')]//*

Will return every element below the header. You could further narrow it down with 

//h2[contains(.,'Headline 1')]//p

for paragraph text, but that would not include the li elements.

Bill Hileman
  • 2,798
  • 2
  • 17
  • 24
  • After taking a second look at the HTML code I used for my test, I realize now that it was inside a . You are obviously correct, thank you for pointing it out. – Bill Hileman Mar 30 '17 at 17:57