Re: Selecting the node which have child without repiti

I see three problems with the current expression, based on your
requirements.  First, you want to filter out those <b> elements that
match preceding <b>s, rather than following <b>s.  Second, it only
looks at siblings, but the second group of <b>s aren't siblings of the
first group.  Last, you really only want to filter out a <b> if the
preceding <b> that it matches has a <c> child, otherwise it would not
have been selected.

Given all that, try something like:

    //b[child::c][not(./@name=(preceding-sibling::b[child::c]/@name |
../preceding-sibling::a/b[child::c]/@name))]

-Brandon :)


On Thu, Nov 18, 2010 at 12:35 PM, Rashi Bhardwaj
<rashi.bhardwaj@xxxxxxxxx> wrote:
> Hi all,
>
> I am facing a problem while selecting a node through xpath....please help me
out
>
> Here is the input xml
>
> <test>
>   <a>
>    <b name ='1'></b>
>    <b name ='2'></b>
>    <b name ='3'>
>         <c></c>
>    </b>
>    <b name ='4'>
>        <c></c>
>        <c></c>
>    </b>
>  </a>
>  <a>
>   <b name ='1'></b>
>   <b name ='2'>
>      <c></c>
>   </b>
>   <b name ='3'></b>
>   <b name ='4'>
>     <c></c>
>   </b>
>  </a>
> </test>
>
> Xpath:
> //b[child::c][not(./@name=following-sibling::b/@name)]
>
> Output:
> b name="3"
> b name="4"
> b name="2"
> b name="4"
>
> What I want is to select all 'b' that have child 'c', provide the 'b'
> with same '@name' which are selected first should not be selected
> further.
>
> Desired output:
>
> Output:
> b name="3"
> b name="4"
> b name="2"
>
> As b with @name = 4 has already been searched, it should not come
> again in search.
>
> Please help me .....thanks a lott....
>
> Rashi