Re: keys with Muenchian Method of grouping help

Hi Jinesh,

> But this stupidly inefficient and bad style especially because I
> have around more than 500,000 author and editor nodes. and preceding
> siblings axis is taking like hours.... An expert suggested me to use
> keys. I was trying to code them:
>
> <xsl:key name="person" match="//publication/author or //publication/editor" use="author|editor" />

The match attribute needs to take a pattern (like the match attribute
of <xsl:template>) rather than an expression. Also, the use attribute
should give you a unique value for each node that's matched by the
match attribute. Since <author> and <editor> elements don't themselves
have <author> or <editor> children, your key above wouldn't find any
value for the <author> and <editor> elements that you're matching.

Try:

<xsl:key name="person"
         match="publication/author | publication/editor"
         use="." />

> <xsl:apply-templates select="*[(author and
> generate-id(.)=generate-id(key('person', author|editor)) or (editor
> and generate-id(.)=generate-id(key('person', author|editor)))]" />

For this, you need to select <author> and <editor> elements that have
unique values. Again, remember that the <author> and <editor> elements
don't have <author> or <editor> children. Try:

<xsl:apply-templates
  select="(//publication/author | //publication/editor)
            [generate-id(.) = generate-id(key('person', .)[1])]" />

Cheers,

Jeni

---
Jeni Tennison
http://www.jenitennison.com/


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list