You need to be careful not to get tripped up by the text nodes in your
document that consist of nothing but whitespace.  Try:

ancestor-or-self::*/text()[normalize-space(.) != ''][last()]

-Brandon :)


On Fri, Apr 15, 2011 at 12:29 PM, Philipp Kursawe
<phil.kursawe@xxxxxxxxx> wrote:
> and I guess to fetch the text() of the current node or its ancestor I
> would have to use:
> ancestor-or-self::*[text/text()][1]/text()
>
> This however does return an empty node for this xml:
> <phrase id="goodbye">
>    <text xml:lang="en">See you tomorrow!</text>
>
>    <text weight="dayOfWeek"
>          days="fri">
>      <text xml:lang="en">Have a nice weekend!</text>
>      <text id="weekend_summer"
>            weight="month"
>            months="4,5,6,7,8,9,10">
>        <text alt="garden" icon="data:binary;..."/>
>        <text alt="boat" icon="data:binary;..."/>
>      </text>
>    </text>
>  </phrase>
>
> and
>
> ancestor-or-self::*[text/text()][1]/text does return the text node.
> But I would like to select the text() in one select.
>
> I tried
> //*[@id='weekend_summer']/ancestor-or-self::*[text/text()][1]/text
> which returns
> <text xml:lang="en">Have a nice weekend!</text>
>
> But i would like to fetch the text() node directly.
>
> Thanks all for your help!
> Phil
> On Fri, Apr 15, 2011 at 8:33 AM, Brandon Ibach
> <brandon.ibach@xxxxxxxxxxxxxxxxxxx> wrote:
>> ancestor-or-self::*[@data][1]/@data
>>
>> -Brandon :)
>>
>>
>> On Fri, Apr 15, 2011 at 2:06 AM, Philipp Kursawe <phil.kursawe@xxxxxxxxx>
wrote:
>>> Hello,
>>>
>>> given the following xml:
>>>
>>> <node data="foo">
>>>   <text id="1" data="bar"/>
>>>   <text id="2"/>
>>> </node>
>>>
>>> I would like to know the XPath expression to select @data which
>>> results in "bar" for @id=1 and "foo" for @id=2.
>>>
>>> The general idea is to search for the attribute in the anchestor if
>>> the current node does not contain the attribute all the way up to the
>>> root node.
>>>
>>> Thanks for your help!
>>>
>>> Phil