Filter elements by attribute values - xslt

Question

I'd like to filter all elements with a duplicate attribute value.
XML:
<?xml version="1.0" encoding="UTF-8"?>
<elements>
<element pos="1"/>
<element pos="2"/>
<element pos="2"/>
<element pos="3"/>
</elements>
XSLT:
<?xml version="1.0" encoding="UTF-8"?>
<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
xmlns:xs="http://www.w3.org/2001/XMLSchema" exclude-result-prefixes="xs" version="2.0">
<xsl:output indent="yes"/>
<xsl:template match="/elements/element">
<xsl:if test="#pos != preceding-sibling::element/#pos">
<xsl:copy-of select="."/>
</xsl:if>
</xsl:template>
<xsl:template match="text() | #*"/>
</xsl:stylesheet>
Output:
<?xml version="1.0" encoding="UTF-8"?>
<element pos="2"/>
<element pos="2"/>
<element pos="3"/>
Output (should be):
<?xml version="1.0" encoding="UTF-8"?>
<element pos="1"/>
<element pos="2"/>
<element pos="3"/>
What am I doing wrong? I don't get it. :-(

Answer 1

I would probably use an XSLT key to group the elements and identify duplicates. Perhaps something like this:
<xsl:key name="elements-by-pos" match="element" use="#pos"/>
<xsl:template match="elements/element">
<xsl:copy-of select="
.[generate-id(.) = generate-id(key('elements-by-pos', #pos)[1])]
"/>
</xsl:template>
But as to why your code doesn't work as you expect, I think it's because this XPath expression doesn't mean what you think it does:
#pos != preceding-sibling::element/#pos
This expression does not mean "are there no previous #pos which are equal to the current #pos?"; rather it means "are there any previous #pos which are unequal to the current #pos?"
You are using the != operator to compare a single #pos attribute with the set of #pos attributes which appear earlier in the document. The operator will return a true value if any member of the set satisfies the condition, i.e. if there is at least one preceding #pos attribute which is not equal to the current #pos. See the section on Boolean expressions in the XPath 1.0 spec for details.
Walking through your example:
For the first #pos (whose value is 1), there are no preceding
#pos attributes, so there are none which are unequal to the current
#pos.
For the second #pos (2) there is a previous #pos whose
value is unequal (i.e. the first #pos, whose value is 1).
The third #pos (whose value is also 2) will be copied because the first #pos 1 again satisfies the != condition.
The final #pos (3) appears because any of the
previous #pos attributes will satisfy the condition.
If you were to change that expression to not(#pos = preceding-sibling::element/#pos) then you would see the behaviour you want.