Flatten XML tree using XSLT - optional but repeating nodes

Question

I'm currently trying to flatten a xml structure to display in a simple table. The basic problem is, that the xml contains repeating nodes on various levels - each combination of the extracted nodes should result in a separate output node.

The xml document looks like this:

<customer>
    <name>Mustermann</name>
    <contract>
        <contract_id>C1</contract_id>
        <products>
            <product>
                <name>Product C1.P1</name>
                <price>23.12</price>
                <properties>
                    <property>Property C1.P1.A</property>
                    <property>Property C1.P1.B</property>
                </properties>
            </product>
            <product>
                <name>Product C1.P2</name>
                <price>2.32</price>
            </product>
        </products>
    </contract>
    <contract>
        <contract_id>C2</contract_id>
        <products>
            <product>
                <name>Product C2.P1</name>
                <price>143.33</price>
            </product>
            <product>
                <name>Product C2.P2</name>
                <price>231.76</price>
                <properties>
                    <property>Property C2.P2.A</property>
                    <property>Property C2.P2.B</property>
                </properties>
            </product>
        </products>
    </contract>
    <contract>
        <contract_id>C3</contract_id>
    </contract>
</customer>

So a contract does not need to have products, a product does not need to have properties.

The result should flatten the repeating nodes and form a separate result node for each set of extracted nodes (which of course introduces some redundancy in the result).

The output should look like this:

<output>
    <data>
        <customer_name>Mustermann</customer_name>
        <contract_id>C1</contract_id>
        <product_name>Product C1.P1</product_name>
        <product_price>23.12</product_price>
        <product_property>Property C1.P1.A</product_property>
    </data>
    <data>
        <customer_name>Mustermann</customer_name>
        <contract_id>C1</contract_id>
        <product_name>Product C1.P1</product_name>
        <product_price>23.12</product_price>
        <product_property>Property C1.P1.B</product_property>
    </data>
    <data>
        <customer_name>Mustermann</customer_name>
        <contract_id>C1</contract_id>
        <product_name>Product C1.P2</product_name>
        <product_price>2.32</product_price>
        <product_property/>
    </data>
    <data>
        <customer_name>Mustermann</customer_name>
        <contract_id>C2</contract_id>
        <product_name>Product C2.P1</product_name>
        <product_price>143.33</product_price>
        <product_property/>
    </data>
    <data>
        <customer_name>Mustermann</customer_name>
        <contract_id>C2</contract_id>
        <product_name>Product C2.P2</product_name>
        <product_price>231.76</product_price>
        <product_property>Property C2.P2.A</product_property>
    </data>
    <data>
        <customer_name>Mustermann</customer_name>
        <contract_id>C2</contract_id>
        <product_name>Product C2.P2</product_name>
        <product_price>231.76</product_price>
        <product_property>Property C2.P2.B</product_property>
    </data>
    <data>
        <customer_name>Mustermann</customer_name>
        <contract_id>C3</contract_id>
        <product_name/>
        <product_price/>
        <product_property/>
    </data>
</output>

I tried the following xslt stylesheet:

<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
<xsl:template match="/customer">
  <output>
  <xsl:for-each select="contract">
    <xsl:choose>
    <xsl:when test="products/product">
      <xsl:for-each select="products/product">
        <xsl:choose>
          <xsl:when test="properties/property">
            <xsl:for-each select="properties/property">
              <xsl:call-template name="data">
                <xsl:with-param name="customer_name" select="/customer/name"/>
                <xsl:with-param name="contract_id" select="../../../../contract_id"/>
                <xsl:with-param name="product_name" select="../../name"/>
                <xsl:with-param name="product_price" select="../../price"/>
                <xsl:with-param name="property" select="."/>
              </xsl:call-template>
            </xsl:for-each>
          </xsl:when>
          <xsl:otherwise>
             <xsl:call-template name="data">
               <xsl:with-param name="customer_name" select="/customer/name"/>
               <xsl:with-param name="contract_id" select="../../contract_id"/>
               <xsl:with-param name="product_name" select="name"/>
               <xsl:with-param name="product_price" select="price"/>
             </xsl:call-template>
          </xsl:otherwise>    
        </xsl:choose>
      </xsl:for-each>
      </xsl:when>
      <xsl:otherwise>
          <xsl:call-template name="data">
            <xsl:with-param name="customer_name" select="/customer/name"/>
            <xsl:with-param name="contract_id" select="contract_id"/>
            </xsl:call-template>
      </xsl:otherwise>    
    </xsl:choose>
  </xsl:for-each>
  </output>
</xsl:template>
<xsl:template name="data">
<xsl:param name="customer_name"/>
<xsl:param name="contract_id"/>
<xsl:param name="product_name"/>
<xsl:param name="product_price"/>
<xsl:param name="property"/>
  <data>
    <customer_name><xsl:value-of select="$customer_name"/></customer_name>
    <contract_id><xsl:value-of select="$contract_id"/></contract_id>
    <product_name><xsl:value-of select="$product_name"/></product_name>
    <product_price><xsl:value-of select="$product_price"/></product_price>
    <product_property><xsl:value-of select="$property"/></product_property>
  </data>
</xsl:template>
</xsl:stylesheet>

This produces the desired result, but to me, this looks rather ugly. The naive approach to just do a nested loop (loop over contract, loop over products inside, and so on) does not work, because it does not pick up all output nodes (e.g. a contract without product is ommited).

Background: I'd like to use the xslt stylesheet in a oracle xmltable query, and the intermediate result will be mapped to a simple row/column output. Therefor each result needs to be in its own row.

Any help is appreciated!

Answer 1

This produces the desired result, but to me, this looks rather ugly.

Well, it's an ugly problem. Is the result really useful for anything? I would have thought a separate table (and hence a separate stylesheet) for each level would be more practical.

Still, if that's the result you want, you may try a more elegant way to produce it:

XSLT 1.0

<xsl:stylesheet version="1.0" 
xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
<xsl:output method="xml" version="1.0" encoding="UTF-8" indent="yes"/>
<xsl:strip-space elements="*"/>

<xsl:template match="/customer">
    <output>
        <xsl:apply-templates/>
    </output>
</xsl:template>

<xsl:template match="contract[not(products)] | product[not(properties)] | property">
    <data>
        <customer_name><xsl:value-of select="ancestor::customer/name"/></customer_name>
        <contract_id><xsl:value-of select="ancestor-or-self::contract/contract_id"/></contract_id>
        <product_name><xsl:value-of select="ancestor-or-self::product/name"/></product_name>
        <product_price><xsl:value-of select="ancestor-or-self::product/price"/></product_price>
        <product_property><xsl:value-of select="self::property"/></product_property>
    </data>
</xsl:template>

<xsl:template match="text()"/>

</xsl:stylesheet>