XSLT: Combine elements without duplicated

Question

I want to do a transformation from XML to text combining some elements, but avoiding duplicates in the output. The XML would be something like that:

<A>
  <B>
    <param1>value0</param1>
    <param2>value1</param2>
  </B>
  <B>
    <param1>value2</param1>
    <param2>value3</param2>
  </B>
  <C>
    <param3>valueC1</param3>
    <D>
      <param4>value0</param4>
      <param5>value4</param5>
    </D>
    <D>
      <param4>value0</param4>
      <param5>value5</param5>
    </D>
    <D>
      <param4>value2</param4>
      <param5>value6</param5>
    </D>
  </C>
  <C>
    <param3>valueC2</param3>
    <D>
      <param4>value0</param4>
      <param5>value5</param5>
    </D>
  </C>
</A>

And the output:

OBJECT: param1=value0, param2=value1, param3=valueC1, param4=value0;
OBJECT: param1=value2, param2=value3, param3=valueC1, param4=value2;
OBJECT: param1=value0, param2=value1, param3=valueC2, param4=value0;

Notes:

• For every D object, look for a match with B objects using D.param4 = B.param1
• If there are two or more D objects into the same C and matching with the same B, print only one of them (in the example, nothing is done with the second D object because it would produce the same line that the first one)
• If there are two D objects matching with the same B, but into different C's, print both (third row in the output example)

I looked for some similar question, but I couldn't find any in the same case.

I guess that it could be done by using keys, but it's too complex.

Thanks!

Regards, Ale.

PS: Sorry for my English.

Answer 1

Given that you're not making any use of param5 in your output it would appear to be possible to simplify the problem to

• for each C
  • find all distinct B elements whose param1 matches the param4 of any of the contained Ds
  • for each of those
    • extract B/param1, B/param2, currentC/param3, B/param1 again (but labelled param4)

This is one way to achieve that using templates.

<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">
  <xsl:output method="text" />
  <xsl:key name="BbyParam1" match="B" use="param1" />

  <xsl:template match="/">
    <xsl:apply-templates select="A/C" />
  </xsl:template>

  <xsl:template match="C">
    <xsl:apply-templates select="key('BbyParam1', D/param4)">
      <xsl:with-param name="currentC" select="." />
    </xsl:apply-templates>
  </xsl:template>

  <xsl:template match="B">
    <xsl:param name="currentC" />
    <xsl:text>OBJECT: param1=</xsl:text>
    <xsl:value-of select="param1" />
    <xsl:text>, param2=</xsl:text>
    <xsl:value-of select="param2" />
    <xsl:text>, param3=</xsl:text>
    <xsl:value-of select="$currentC/param3" />
    <xsl:text>, param4=</xsl:text>
    <xsl:value-of select="param1" />
    <xsl:text>&#10;</xsl:text>
  </xsl:template>
</xsl:stylesheet>

"find all distinct B elements whose param1 matches the param4 of any of the contained Ds" is actually very straightforward due to the fact that when you pass a node set as the second argument to the key function it does precisely this - it returns the set of all nodes whose key value is the string value of any of the nodes in the argument node set, and the returned node set (being a set) is guaranteed to contain no duplicates.

Answer 2

This should do the trick:

<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
  <xsl:output method="text" indent="yes"/>
  <xsl:key name="kB" match="B" use="param1" />
  <xsl:key name="kD" match="D" 
           use="concat(param4, '+', generate-id(..))"/>

  <xsl:template match="/">
    <xsl:apply-templates select="A/C" />
  </xsl:template>

  <xsl:template match="C">
    <xsl:apply-templates select="D[key('kB', param4) and 
                                    generate-id() = 
                                    generate-id(key('kD', 
                                       concat(
                                          param4, 
                                          '+', 
                                          generate-id(..)
                                       )
                                    )[1])]" />
  </xsl:template>

  <xsl:template match="D">
    <xsl:value-of 
      select="concat('OBJECT: param1=', 
                     key('kB', param4)/param1, 
                     ', param2=', 
                     key('kB', param4)/param2,
                     ', param3=',
                     ../param3,
                     ', param4=',
                     param4,
                     '&#xA;')"/>
  </xsl:template>
</xsl:stylesheet>

Output when run on your sample:

OBJECT: param1=value0, param2=value1, param3=valueC1, param4=value0
OBJECT: param1=value2, param2=value3, param3=valueC1, param4=value2
OBJECT: param1=value0, param2=value1, param3=valueC2, param4=value0