Re: [xsl] Finding the current node type

[Kamal Bhatt <kbhatt@xxxxxxxxx>]

If I understand what you are trying to do, you could use local-name(). 
This will give you the name of the node.


Mark Anderson wrote:
> Hi All
>
> I'm trying to figure out a way to determine the "type" of the current
> node in an nodeset.
>
> I have a union of two dissimilar nodesets and I need to number the
> combined set sequentially in my output. I also need to output details
> from each node, but as they are completely different I need to process
> each node type differently based on it's contents.
>
> Here's a silly example
>
> <products>
> 	<speaker>
> 		<woofer>12 inch</woofer>
> 		<rms_power>70 watts</rms_power>
> 		<peak_power>95 watts</peak_power>
> 		<height>40 inches</height>
> 		<width>18 inches</width>
> 		<depth>18 inches</depth>
> 	</speaker>
> 	<laptop>
> 		<disk>80GB</disk>
> 		<screen>15 inch</screen>
> 		<cpu>Intel Core Duo 2.3ghz</cpu>
> 		<weight>5.2lbs</weight>
> 	</laptop>
> 	...
> </products>
>
> So, I need to do something like:
>
> <xsl:for-each select="speaker | laptop">
> 	<!-- Assign a sequential Number -->
> 	Item No: <xsl:value-of select="position()"/> <br />
> 	<!-- Now output item-specific details -->
> 	<xsl:choose>
> 		<xsl:when test="node = speaker">
> 			RMS Power: <xsl:value-of select="rms_power"/>
> <br />
> 			Peak Power: <xsl:value-of select="peak_power"/>
> <br />	
> 		</xsl:when>
> 		<xsl:when test="node = laptop">
> 			Disk Drive: <xsl:value-of select="disk"/> <br />
> 			CPU: <xsl:value-of select="cpu"/> <br />
>
> 		</xsl:when>	
> 	</xsl:choose>
> </xsl:for-each> 
>
> So, what I need to figure out is the test in the when clause to see
> which type of node I'm processing.
>
> I could hack it with something like:
> 	
> <xsl:when test="count(rms_power) &gt; 0"> 
> 	<!-- Must be a speaker -->
> 	RMS Power: <xsl:value-of select="rms_power"/>
> 	Peak Power: <xsl:value-of select="peak_power"/>		
> </xsl:when>
>
> However, it's messy and I have to find unique child nodes for every
> product.
>
> I'm sure there's a more elegant way
>
> Can anyone help?
>
> TIA
>
> Mark
>
>
>   


--
Kamal Bhatt