RE: [xsl] Filtering/Removing Truly Duplicate Elements

["Michael Kay" <mike@xxxxxxxxxxxx>]

Removing duplicates is essentially the same problem as grouping (you group
the elements and then output the first one in each group). So for XSLT 2.0
you can use xsl:for-each-group, or if you really have a very good reason to
stick with 1.0 you can use Muenchian grouping, see
http://www.jenitennison.com/xslt/grouping.

In 2.0 it's

<xsl:for-each-group select="//xs:element" group-by="concat(@name, '|',
@type)">
  <xs:element name="{current-group()[1]/@name}" 
              type="{current-group()[1]/@type}"/> 
</xsl:for-each-group>

Michael Kay
http://www.saxonica.com/ 

> -----Original Message-----
> From: Wasiq Shaikh [mailto:wasiq911@xxxxxxxxxxx] 
> Sent: 18 July 2007 19:36
> To: xsl-list@xxxxxxxxxxxxxxxxxxxxxx
> Subject: [xsl] Filtering/Removing Truly Duplicate Elements
> 
> Hello everyone,
> 
> I have searched everywhere for a similar problem but most 
> problems are actually very simple with simple solutions. But 
> the techniques are the same. 
> Except I cant seem to put differenct peices together. When i 
> run my XSL, the desired result is not desired! :)
> 
> The Objective:
> Remove truly duplicate elements, not just by name but by its 
> attributes and attribute values.
> 
> The Example:
> I have an xml input file (which is actually an XSD) with 
> various <xsd:element>'s which may have many attributes.
> The attributes I care about are "name" and "type". Below is a 
> sample input
> file:
> 
> <xsd:element name="ABC" type="123" ... >
>    <xsd:element name="def" type="456" .../>
>    <xsd:element name="ghi" type="456" .../> </xsd:element>
> 
> <xsd:element name="ABC" type="456" ...>
>    <xsd:element name="ghi">
>       <xsd:element name="def" type="456" .../>
>       <xsd:element name="def"/>
>    </xsd:element>
>    <xsd:element name="ABC"/>
>    <xsd:element name="ABC"/>
>    <xsd:element name="ghi" type="456" .../> </xsd:element>
> 
> The desired result should be a flat list of elements:
> 
> <uniqueElements>
>    <element name="ABC"/>
>    <element name="ABC" type="123"/>
>    <element name="ABC" type="456"/>
>    <element name="def"/>
>    <element name="def" type="456"/>
>    <element name="ghi"/>
>    <element name="ghi" type="456"/>
> </uniqueElements>
> 
> The order of the result doesn't matter. But understand that 
> all those elements are distinctly unique. I'm using Xalan 
> Java 2.7/XPath 1.0.
> 
> My XSL Logic:
> Since i need a flat list i will interate through "//xsd:element".
> So, for every element encountered, check to see if there are 
> any elements before it with the same name, and same type.
> 
> Sounds simple enough! But it isn't. I cant get the right 
> combination of conditional logic with correct XPath syntax to 
> execute this task. Maybe this task can not be done at all? I 
> would like to know first if I'm wasting my time trying to 
> figure this out or if there is any hope.
> 
> Please help as I'm sure this will benefit anyone else wanting 
> to do the same task. Thank you!
> 
> Wasiq Shaikh
> 
> _________________________________________________________________
> Windows Live Hotmail. Even hotter than before. Get a better look now. 
> www.newhotmail.ca?icid=WLHMENCA148