Xpath - choose A if B is 0 else choose B - xpath

Question

I'm trying to gather some informations from an estatepage and i've run into this logic problem before and now i'm asking for your help.
basicly it's there are two different ways the table with info i'm looking at
<table>
<tr>
<th>Bolig areal</th> (estate size)
<td>0 m</td>
</tr>
<tr>
<th>Grund areal</th> (ground size)
<td>5000 m</td>
</tr>
</table>
and
<table>
<tr>
<th>Bolig areal</th> (estate size)
<td>123 m</td>
</tr>
<tr>
<th>Grund areal</th> (ground size)
<td>5000 m</td>
</tr>
</table>
I want to get the 'bolig areal'(estate size). But if 'bolig areal' is 0 m, then take Grund areal(ground size)
I'm currently stuck on an Xpath expression like:
//div[#class='description-element description-info shadow']//th[contains(.,'Grund areal') or contains(.,'Bolig areal')]/following-sibling::td/text()[not(contains(.,'0 m'))]
Where I select both the <th>, selects the following <td>/text, where the text is != 0m.
Which selects the Ground area.
How do I add the "if 'Bolig areal' is different from 0 m pick 'Bolig areal', otherwise use 'Ground areal'
solution:
".//tr[th = 'Bolig areal' and td != '0 m' or th = 'Grund areal' and not(ancestor::table/tr[th = 'Bolig areal' and td != '0 m'])]/td/text()"

Answer 1

try this:
//tr[th = 'Bolig areal' and td != '0 m' or th = 'Grund areal' and not(ancestor::table/tr[th = 'Bolig areal' and td != '0 m'])]/td
This selects the tr that has either th = 'Bolig areal' and td != '0 m' or the tr that has th = 'Grund areal' if the first expression th = 'Bolig areal' and td != '0 m' is not true for the current table. From that selected tr the td is selected.

Answer 2

What about:
//table[tr[th="Bolig areal"][td != "0 m"]]/tr[th="Bolig areal"]/td/text()
|
//table[tr[th="Bolig areal"][td = "0 m"]]/tr[th="Grund areal"]/td/text()
This is essentially a translation of:
if a table contains a row with a "Bolig areal" th that has a value (td) different from "0 m", pick that value (td/text())
alternatively (| operator), if the "Bolig areal" value is "0 m", select the value from the row with header "Grund areal"
You could use contains(th, "Bolig areal") instead of th="Bolig areal"
Example Python session with lxml (libxml2):
In [1]: import lxml.html
In [2]: doc1 = lxml.html.fromstring("""<table>
...: <tr>
...: <th>Bolig areal</th> (estate size)
...: <td>0 m</td>
...: </tr>
...: <tr>
...: <th>Grund areal</th> (ground size)
...: <td>5000 m</td>
...: </tr>
...: </table>""")
In [3]: doc2 = lxml.html.fromstring("""<table>
...: <tr>
...: <th>Bolig areal</th> (estate size)
...: <td>123 m</td>
...: </tr>
...: <tr>
...: <th>Grund areal</th> (ground size)
...: <td>5000 m</td>
...: </tr>
...: </table>""")
In [4]: xp = lxml.etree.XPath("""//table[tr[th="Bolig areal"][td != "0 m"]]/tr[th="Bolig areal"]/td/text()
...: |
...: //table[tr[th="Bolig areal"][td = "0 m"]]/tr[th="Grund areal"]/td/text()""")
In [5]: xp(doc1)
Out[5]: ['5000 m']
In [6]: xp(doc2)
Out[6]: ['123 m']
In [7]:

Answer 3

You want the first th that is not followed by 0 m, then extract the first following td:
//th[(.='Bolig areal' and following-sibling::td!='0 m') or (.='Grund areal' and ../..//th[.='Bolig areal']/following-sibling::td='0 m')]/following-sibling::td[1]