Well, it should be in the same namespace as count (that is, in the default
fn: namespace).
"
The default prefix for the function namespace is fn:, and the URI is:
http://www.w3.org/2005/02/xpath-functions.
"

However, I see you already have a working solution...

michael

-----Original Message-----
From: Mark Peters [mailto:markpeters.work@xxxxxxxxx]
Sent: 4. jzna 2009 21:07
To: xsl-list@xxxxxxxxxxxxxxxxxxxxxx
Subject: Re:  Determining whether no single instance of a specific node
exists

Hi Mtekel,

I'm encountering the following error:

XSLT Error (javax.xml.transform.TransformerException): Error in
expression not(exists( (b/c) ) ): Unknown system function: exists

Do I need to specify the URI for the XSLT namespace somewhere?

Thanks,
Mark


On Thu, Jun 4, 2009 at 2:51 PM, Mtekel <thx@xxxxxxxxx> wrote:
> Hello,
>
> <xsl: if test="not(exists( (b/c) ) )" > ... </xsl:if>
>
> Check other functions here:
> http://www.w3schools.com/xpath/xpath_functions.asp#sequence
>
> You can use all xpath functions in the if/when test condition.
>
> m.
>
> -----Original Message-----
> From: Mark Peters [mailto:markpeters.work@xxxxxxxxx]
> Sent: 4. jzna 2009 20:21
> To: xsl-list@xxxxxxxxxxxxxxxxxxxxxx
> Subject:  Determining whether no single instance of a specific node
> exists
>
> Hi Everyone,
>
> I have the following XML document:
>
>     <a>
>             <b/>
>             <b/>
>             <b/>
>     </a>
>     <a>
>             <b/>
>             <b/>
>             <b>
>                     <c/>
>             </b>
>
>     </a>
>     <a>
>             <b/>
>             <b/>
>             <b/>
>     </a>
>
> For each <b/> set, I want to apply a transformation only if no single
> instance of <c/> exists. I know how to check whether a node does
> exist, but not how to check whether no single instance of the node
> exists.
>
> I've tried the following conditional statements without success:
>
> 1,
> <xsl:when test="not(a/b/c)">
>
> 2.
> <xsl:when test="count(not(a/b/c)) &gt;= 1">
>
>
> What conditional statement should I use?
>
> Thanks in advance for any help.
>
> Mark