Question

I have a XML I am wanting to transform it with xslt.
The input looks like this.

<xmeml>
<Doc>
    <Test>
        <Unit>abc</Unit>
        <Unit2>1234</Unit2>
        <Unit3>tuvw</Unit3>
    </Test>
    <Test>
        <Unit>bcd</Unit>
        <Unit2>2345</Unit2>
        <Unit3>wxyz</Unit3>
    </Test>
</Doc>
<Doc>
    <Test>
        <Unit>abc</Unit>
        <Unit2>3456</Unit2>
        <Unit3>wxyz</Unit3>
    </Test>
    <Test>
        <Unit>cde</Unit>
        <Unit2>3456</Unit2>
        <Unit3>wxyz</Unit3>
    </Test> 
</Doc>
<Doc>
    <Test>
        <Unit>abc</Unit>
        <Unit2>1234</Unit2>
        <Unit3>wxyz</Unit3>
    </Test>
    <Test>
        <Unit>def</Unit>
        <Unit2>4567</Unit2>
        <Unit3>wxyz</Unit3>
    </Test> 
</Doc>
<Doc>
    <Test>
        <Unit>abc</Unit>
        <Unit2>1234</Unit2>
        <Unit3>uvwx</Unit3>
    </Test>
    <Test>
        <Unit>efg</Unit>
        <Unit2>2345</Unit2>
        <Unit3>wxyz</Unit3>
    </Test> 
</Doc>
</xmeml>

the output should look like this.

<xmeml>
<Doc>
    <Test>
        <Unit>bcd</Unit>
        <Unit2>2345</Unit2>
        <Unit3>wxyz</Unit3>
    </Test>
</Doc>
<Doc>
    <Test>
        <Unit>abc</Unit>
        <Unit2>3456</Unit2>
        <Unit3>wxyz</Unit3>
    </Test>
    <Test>
        <Unit>cde</Unit>
        <Unit2>3456</Unit2>
        <Unit3>wxyz</Unit3>
    </Test> 
</Doc>
<Doc>
    <Test>
        <Unit>abc</Unit>
        <Unit2>1234</Unit2>
        <Unit3>wxyz</Unit3>
    </Test>
    <Test>
        <Unit>def</Unit>
        <Unit2>4567</Unit2>
        <Unit3>wxyz</Unit3>
    </Test> 
</Doc>
<Doc>
    <Test>
        <Unit>efg</Unit>
        <Unit2>2345</Unit2>
        <Unit3>wxyz</Unit3>
    </Test> 
</Doc>
</xmeml>

I am wanting to remove any Test nodes that match the following criteria.
– The Unit3 child node starts-with tuv or uvw.
– AND Where both the Unit AND Unit2 vales are both found repeated/duplicated in another Test node

Thanks in advance for your help.

Answer 1

How about this tweak of Dimitre’s solution …

<xsl:stylesheet version="1.0"
 xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
 <xsl:output omit-xml-declaration="yes" indent="yes"/>
 <xsl:strip-space elements="*"/>
 <xsl:key name="kTest" match="Test" use="concat(Unit, '+', Unit2)"/>

 <xsl:template match="node()|@*">
  <xsl:copy>
   <xsl:apply-templates select="node()|@*"/>
  </xsl:copy>
 </xsl:template>

 <xsl:template match=
  "Test[key('kTest', concat(Unit, '+', Unit2))[2]]
   [starts-with(Unit3, 'tuv') or starts-with(Unit3, 'uvw')]"/>
</xsl:stylesheet>

As I understand it, the rules of transformation are:

1. Copy all nodes as is, except as indicated following.
2. Exclude all Test nodes have that the following properties:

2.1 There exists another Test node (call it Test(other)) in the document such that:

  (Test(reference) / Unit = Test(other) / Unit)   AND
  (Test(reference) / Unit2 = Test(other) / Unit2)

2.2 The Unit3 child of the reference Test node starts with tuv or uvw

The output for this is exactly as the OP posted. Interestly, if I have understood the problem correctly, Dimitre has just recently provided a solution to almost the same problem as yours here XSL comparison of nodes . Again, if I have understood your problem correctly, the title is misleading. “Removing duplicates” is generally taken to mean de-duping, as in (1,2,2,3) –> (1,2,3) ; whereas what you want, like (1,2,2,3) –> (1,3) is generally described as “selecting unique nodes/values”.