From: "Jesús Gabriel y Galán" <jgabrielygalan@...>
Date: 2011-11-25T00:32:32+09:00
Subject: Re: NOKIGIRI XPATH

On Thu, Nov 24, 2011 at 1:49 PM, Robert Klemme
<shortcutter@googlemail.com> wrote:
> 2011/11/24 Jesús Gabriel y Galán <jgabrielygalan@gmail.com>:
>> On Thu, Nov 24, 2011 at 12:30 PM, Ruby Mania <prateek123@gmail.com> wrote:
>>> Thanks a lot for help. But it matched CDEF and all nodes after that even
>>> if key != english
>>
>> I'm not sure why is this. I'm still trying to come up with a good
>> XPath that will return just that node,
>
> Well, there could be many matches and from the original posting I
> cannot see that only the first is needed.

What I don't understand is why that xpath returns nodes whose
preceding key sibling doesn't have 'English' as value.
I mean:


<topics>
  <topic>
    <key>English</key>
    <topic><data>CDEF</data></topic>
  </topic>
  <topic>
    <key>Spanish</key>
    <topic><data>ABC</data></topic>
  </topic>
</topics>

Why that xpath returns the ABC also. I would have thought that
following-sibling for <key>English</key> would only be the
<topic><data>CDEF</data></topic>, from which we are selecting the data
node.


>> but in the meantime you can do this:
>>
>> doc.xpath("//key[. = 'English']/following-sibling::topic/data")[0]
>
> Or better
>
> doc.at_xpath("//key[. = 'English']/following-sibling::topic[1]/data")
>
> I would probably do
>
> doc.xpath('//topic[preceding-sibling::key[text()="English"]]//data')
>
> or, for one hit only
>
> doc.at_xpath('//topic[preceding-sibling::key[text()="English"]][1]//data')
>
> Not sure about efficiency but I prefer it visually to have the path to
> the selected node as basis and use criteria in [] for filtering.

I agree with you, and I would guess this is more efficient, since
nokogiri doesn't have to return as many nodes.

Jesus.