Re: [xsl] Got an XSLT function that does value intersection (notidentity intersection)?

["Roger L Costello costello@xxxxxxxxx" <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx>]

Michael Kay wrote:

> In this situation, you can use grouping:
>
> <xsl:for-each-group select="row" group-by="x, y" composite="yes">
>     <xsl:sequence select="current-group()[1]"/>
> </xsl:for-each>

That is a beautiful solution.

Just curious, what is the time complexity of that loop in Saxon? Is Saxon able
to do in in better than O(n**2)?

/Roger

-----Original Message-----
From: Michael Kay mike@xxxxxxxxxxxx <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx>
Sent: Thursday, July 8, 2021 9:32 AM
To: xsl-list <xsl-list@xxxxxxxxxxxxxxxxxxxxxx>
Subject: [EXT] Re: [xsl] Got an XSLT function that does value intersection
(notidentity intersection)?

In this situation, you can use grouping:

<xsl:for-each-group select="row" group-by="x, y" composite="yes">
  <xsl:sequence select="current-group()[1]"/>
</xsl:for-each>

If your comparison operation is more complex (e.g., deep-equal()) then it gets
more difficult, and the best way might be to form the canonical serialization
of each element and group on that. (Saxon offers canonical serialization with
the extension attribute saxon:canonical="yes").

Michael Kay
Saxonica

> On 8 Jul 2021, at 14:19, Roger L Costello costello@xxxxxxxxx
<xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx> wrote:
>
> Hi Folks,
>
> I have an XML document that consists of <row> elements:
>
> <Document>
>    <row>
>        <x>1</x>
>       <y>2</y>
>    </row>
>    <row>
>        <x>3</x>
>       <y>4</y>
>    </row>
>    <row>
>        <x>1</x>
>       <y>2</y>
>    </row>
> </Document>
>
> I want to eliminate duplicate rows.
>
> The XPath intersect operator is appealing:
>
> $document intersect $document
>
> because intersect eliminates duplicates. Unfortunately, that doesn't work
because row1 and row2 are "duplicates" only if they are the exact same <row>
element, i.e., same identity.
>
> I need an intersection operator that eliminates duplicates based on value.
Since row[1] and row[3] above have the same elements with the same values,
they are duplicates.
>
> Do you have a function that does intersection based on value, not identity?
Or, perhaps you recommend another way to eliminate duplicates?
>
> /Roger