Hi,
I've a structure like that:

   <ProgTools>
      <ProgTool>
         <Keyword>
            <Characteristic>PROGNAME</Characteristic>
            <ValueDescription>bla1</ValueDescription>
         </Keyword>
         <Keyword>
            <Characteristic>PROGTYPE</Characteristic>
            <ValueDescription>bla2</ValueDescription>
         </Keyword>
         ...

NOT all the keywords will be given.
A record ProgTool may contain cell 1, 2, 3.
The 2nd rec. may contain only 1 and 3.
Another may contain 2 and 3 etc.

The table should be written with <tr> and <td>, empty fields with a space inside (&nbsp;).

Any help?

If you always want characteristic S in column 1, W in column 2, and H in column 3, then do

<tr>
<td>
  <xsl:value-of select="Keyword[Characteristic='S']/ValueDescription"/>
</td>
<td>
  <xsl:value-of select="Keyword[Characteristic='W']/ValueDescription"/>
</td>
<td>
  <xsl:value-of select="Keyword[Characteristic='H']/ValueDescription"/>
</td>
</tr>

If that's not what you want, then you'll have to explain it better.

Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference

Hi Michael,
I try to explain better, sorry for my bad english ;)
I have a XML structure looking like a 'incomplete' 2-dimensional matrix.

l = line, c = column

l1c1, l1c2, l1c3
----, l2c2, l2c2
l3c1, ----, l3c3
l4c1, l4c2, ----

As you see, not all elements of the 'matrix' are given by the XML, just the lNcN values (N=1...n).

I'm now looking for a XSL programming algorithm, which displays all the given elements by its values BUT also displays the '----' values as empty HTM table cells (<td></td>).

I'm using the lastest Saxon engine shipped with Stylus Studio.

Is that any clearer?

Ciao

Please try to show what the input XML looks like, and how this relates to the desired output.

Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference

<ProgTools>
      <ProgTool>
         <Keyword>
            <Characteristic>PROGNAME</Characteristic>
            <ValueDescription>l1c1</ValueDescription>
         </Keyword>
         <Keyword>
            <Characteristic>PROGTYPE</Characteristic>
            <ValueDescription>l1c2</ValueDescription>
         </Keyword>
         <Keyword>
            <Characteristic>PROGADAPTER</Characteristic>
            <ValueDescription>l1c3</ValueDescription>
         </Keyword>
         ...



INPUT:

ProgTools has n ProgTool elements (lines).
Each ProgTool has 1 to n Keywords/Characteristics.

OUTPUT (HTM):

                <table>
                    <tr>
                      <td>l1c1</td>
                      <td>l1c2</td>
                      <td>l1c3</td>
                    </tr>
                    <tr>
                      <td>l2c1</td>
                      <td>&nbsp;</td>
                      <td>l2c3</td>
                    </tr>
                    <tr>
                      <td>l3c1</td>
                      <td>l3c2</td>
                      <td>&nbsp;</td>
                    </tr>
                </table>

As you see from the output, not ALL the fields of one 'record' will have to be filled, so the algorithm should be able to handle 'incomplete' records.

Thanks for your help!

Why can't you do what I ask?

You haven't shown me the data that produces the second and third rows of your table.

If they look like I think they look, then the code I gave you before works. If that code doesn't work, then it must be because I made a wrong guess about your input. That's why I asked you to show me the input.

Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference

Hi Michael,

your solution:

<tr>
<td>
  <xsl:value-of select="Keyword[Characteristic='S']/ValueDescription"/>
</td>
<td>
  <xsl:value-of select="Keyword[Characteristic='W']/ValueDescription"/>
</td>
<td>
  <xsl:value-of select="Keyword[Characteristic='H']/ValueDescription"/>
</td>
</tr>

worked! Thank you!!