This table illustrates several products that are sold in random 
locations along with the various prices for each location.

        loc1  loc2  loc3  loc4
      -------------------------
prod1 | 10.00       12.00
prod2 | 11.00 11.00       10.00
prod3 |       12.00


The XML for this table is:

<price_summ>
   <prod name="prod1">
      <loc name="loc1">
		  <price>10.00</price>
	  </loc>
      <loc name="loc3">
           <price>12.00</price>
		</loc>
   </prod>
   <prod name="prod2">
      <loc name="loc1">
		  <price>11.00</price>
	  </loc>
      <loc name="loc2">
           <price>11.00</price>
	   </loc>
      <loc name="loc4">
           <price>10.00</price>
	   </loc>
   </prod>
   <prod name="prod3">
      <loc name="loc2">
           <price>12.00</price>
	   </loc>
   </prod>
</price_summ>

I can apply a template on <prod> and would like to do so on <loc>, 
but how do I fill-in the location cells that do not have a value.

I've used the algorithm listed below, but THERE HAS TO BE A BETTER WAY!

(BTW, the alorithm below is an example of my procedural thinking.)

Any ideas?
			
	template select="prod"

	<tr>
		<td>value-of select="prod/@name"</td>

	choose 
		when select="loc[@name="loc1"]
			<td>value-of "loc[@name="loc1]/price</td>
		/when
		other
			<td>&nbsp;</td>
		/other
	/choose
	choose 
		when select="loc[@name="loc2"]
			<td>value-of "loc[@name="loc2]/price</td>
		/when
		other
			<td>&nbsp;</td>
		/other
	/choose
	choose 
		when select="loc[@name="loc3"]
			<td>value-of "loc[@name="loc3]/price</td>
		/when
		other
			<td>&nbsp;</td>
		/other
	/choose
	choose 
		when select="loc[@name="loc4"]
			<td>value-of "loc[@name="loc4]/price</td>
		/when
		other
			<td>&nbsp;</td>
		/other
	/choose

	</tr>
	/template