Grouping/Merging

Jens Rudolph · June 5th, 2007, 10:14 AM

Hello,

I have a problem with grouping respectively merging xml elements. I can only use XPath 1.0 and XSL 1.0.

This is the xml structure:

Code:

<Root>
  <Items>
    <Item id="A1" name="Part" version="1.0"/>
    <Item id="A2" name="Part" version="2.0"/>
    <Item id="A3" name="SubPart" version="0.5"/>
    <Item id="A4" name="SubPart" version="1.1"/>
    <Item id="A5" name="ProductA" version="1.5"/>
    <Item id="A6" name="ProductA" version="2.2"/>
    <Item id="A7" name="ProductB" version="2.1"/>
  </Items>
  <Configurations>
    <Configuration id="B7" idItem="A7" installed="false">
      <Parts>
        <Part key="[ProductB]" id="B7" idItem="A7"/>
        <Part key="[Part]" id="B1" idItem="A1"/>
        <Part key="[SubPart]" id="B9" idItem="A3"/>
      </Parts>
    </Configuration>
    <Configuration id="B3" idItem="A5" installed="false">
      <Parts>
        <Part key="[Part]" id="B1" idItem="A1"/>
        <Part key="[ProductA]" id="B3" idItem="A5"/>
        <Part key="[SubPart]" id="B9" idItem="A3"/>
      </Parts>
    </Configuration>
    <Configuration id="B1" idItem="A1" installed="false">
      <Parts>
        <Part key="[Part]" id="B1" idItem="A1"/>
        <Part key="[SubPart]" id="B9" idItem="A3"/>
      </Parts>
    </Configuration>
    <Configuration id="B2" idItem="A2" installed="true">
      <Parts>
        <Part key="[Part]" id="B2" idItem="A2"/>
        <Part key="[SubPart]" id="B10" idItem="A4"/>
      </Parts>
    </Configuration>
    <Configuration id="B4" idItem="A5" installed="true">
      <Parts>
        <Part key="[ProductA]" id="B4" idItem="A5"/>
        <Part key="[Part]" id="B2" idItem="A2"/>
        <Part key="[SubPart]" id="B10" idItem="A4"/>
      </Parts>
    </Configuration>
    <Configuration id="B5" idItem="A6" installed="false">
      <Parts>
        <Part key="[ProductA]" id="B5" idItem="A6"/>
        <Part key="[Part]" id="B1" idItem="A1"/>
        <Part key="[SubPart]" id="B9" idItem="A3"/>
      </Parts>
    </Configuration>
    <Configuration id="B6" idItem="A6" installed="true">
      <Parts>
        <Part key="[ProductA]" id="B6" idItem="A6"/>
        <Part key="[Part]" id="B2" idItem="A2"/>
        <Part key="[SubPart]" id="B10" idItem="A4"/>
      </Parts>
    </Configuration>
    <Configuration id="B8" idItem="A7" installed="true">
      <Parts>
        <Part key="[ProductB]" id="B8" idItem="A7"/>
        <Part key="[Part]" id="B2" idItem="A2"/>
        <Part key="[SubPart]" id="B10" idItem="A4"/>
      </Parts>
    </Configuration>
    <Configuration id="B9" idItem="A3" installed="true">
      <Parts>
        <Part key="[SubPart]" id="B9" idItem="A3"/>
      </Parts>
    </Configuration>
    <Configuration id="B10" idItem="A4" installed="true">
      <Parts>
        <Part key="[SubPart]" id="B10" idItem="A4"/>
      </Parts>
    </Configuration>
  </Configurations>
</Root>

There are seven items and each item has different configurations.

The Part elements in the element Parts are not always in the same order
and the element Parts doesn't contain always the same amount of Part
elements for one item name.

It could be possible that there are double versions, e.g.:
ProductA V1.5 consists of:

ProductA V1.5, Part V1.0 and SubPart V0.5
ProductA V1.5, Part V2.0 and SubPart V1.1

The desired output should be a table for each item name.

Code:

Part
----------------------
|  Part  |  SubPart  |
----------------------
|  1.0   |    0.5    |
----------------------
|  2.0   |    1.1    |
----------------------

ProductA
----------------------------------
|  Part  |  ProductA |  SubPart  |
----------------------------------
|   1.0  |     1.5   |    0.5    |
----------------------------------
|   2.0  |     1.5   |    1.1    |
----------------------------------
|   1.0  |     2.2   |    0.5    |
----------------------------------
|   2.0  |     2.2   |    1.1    |
----------------------------------

ProductB
----------------------------------
|  Part  |  ProductB |  SubPart  |
----------------------------------
|   1.0  |     2.1   |    0.5    |
----------------------------------
|   2.0  |     2.1   |    1.1    |
----------------------------------

SubPart
-------------
|  SubPart  |
-------------
|    0.5    |
-------------
|    1.1    |
-------------

I started with this xsl style sheet to get the superset of the headings (parts) from all configurations of an item name:

Code:

<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform">

  <xsl:key name="name" match="Item" use="@name"/>
  <xsl:key name="headings" match="Parts/*" use="@key"/>

  <xsl:variable name="configs" select="//Root/Configurations/Configuration"/>
  <xsl:variable name="items" select="//Root/Items/Item"/>

  <xsl:template match="/">
    <html>
      <body>
        <xsl:for-each select="./Root/Items/Item[count(.|key('name', @name)[1]) = 1]">
          <xsl:sort select="./@name"/>
          <xsl:value-of select="./@name"/>
          <br/>
          <table>
            <tr>
              <xsl:variable name="name" select="./@name"/>
              <xsl:variable name="idsItems" select="../Item[@name=$name]/@id"/>
              <xsl:variable name="parts" select="$configs[@idItem=$idsItems]/Parts/*"/>



              <xsl:for-each select="$parts[count(.|key('headings', @key)[1])=1]">
                <xsl:sort select="./@key"/>
                <th>
                  <xsl:variable name="selectId" select="./@idItem"/>
                  <xsl:value-of select="$items[@id=$selectId]/@name"/>
                </th>
              </xsl:for-each>

            </tr>
          </table>
          <br/>
        </xsl:for-each>
      </body>
    </html>
  </xsl:template>
</xsl:stylesheet>

The output of this xsl style sheet is:

Code:

Part

ProductA
-------------
|  ProductA |
-------------

ProductB
----------------------------------
|  Part  |  ProductB |  SubPart  |
----------------------------------

SubPart

What is wrong with the loop?

Thanks and regards

Jens

mhkay · June 5th, 2007, 11:45 AM

Thanks for submitting such a clear problem. It's surprisingly rare on this forum.

In the line:

<xsl:for-each select="$parts[count(.|key('headings', @key)[1])=1]">

you're selecting a Part element if it's the first Part with that key in the document. So the first time round $parts consists of four Part elements, with idItem=A1 or A2, and in each case there is another Part with the same key that's earlier in the document and that isn't a member of $parts. (This tends to be where Muenchian grouping breaks down - everything is scoped to the document).

As I get no pleasure from solving grouping problems in XSLT 1.0 when there is newer technology that makes the job dead easy, I'll leave the solution to you.

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

Jens Rudolph · June 6th, 2007, 02:50 AM

Hello Micheal,

thanks for your answer.
Then I will look if I could solve it in another way.

Regards

Jens