Re: [xsl] Which is less expensive group by or select distinct-values

Subject: Re: [xsl] Which is less expensive group by or select distinct-values
From: "Dan Vint dvint@xxxxxxxxx" <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx>
Date: Sat, 16 Jul 2016 19:20:22 -0000
>>>I had thought about using keys, but I don't think they would work. So the rest of the requirement is that I have a single file that includes all the 4,000 topics. I create a summary report of all the topics first (which I could see using keys for), but then I step through the content and need to report the details per topic which is where I couldn't see a way to make use of the keys. So the source file has this structure:

<topic>
     <topic>.....</topic>
     <topic>.....</topic>
     <topic>.....</topic>
...
</topic>



At 04:53 AM 7/16/2016, Tony Graham tgraham@xxxxxxxxxxxxx wrote:
Another option that you could try to see how it affects memory usage [1]:

<xsl:key name="terms" match="term[empty(@keyref)]" use="true()" />

<data type="topicreport" name="WDTermList">
   <xsl:value-of separator=", ">
       <xsl:perform-sort select="distinct-values(key('terms', true()))">
           <xsl:sort select="." />
       </xsl:perform-sort>
   </xsl:value-of>
</data>

(borrowing the xsl:perform-sort idea from Martin Honnen, and assuming that the context node is the document element.)

Regards,


Tony Graham.

--------------------------------------------------------------------------- Danny Vint

Panoramic Photography
http://www.dvint.com

voice: 619-647-5780

Current Thread