Re: [xsl] Understanding Identity Transformations

["Joris Gillis" <roac@xxxxxxxxxx>]

Tempore 23:56:58, die 02/11/2005 AD, hinc in  
xsl-list@xxxxxxxxxxxxxxxxxxxxxx scripsit Karl Stubsjoen <kstubs@xxxxxxxxx>:

> Seems like this is pretty standard:
>
> SAMPLE_001:
> <xsl:template match="node()|@*">
>   <xsl:copy>
>     <xsl:apply-templates select="@*"/>
>     <xsl:apply-templates/>
>   </xsl:copy>
> </xsl:template>
>
> But then I've also seen this:
>
> SAMPLE_002:
> <xsl:template match="/ | @* | node()">
>    <xsl:copy>
>      <xsl:apply-templates select="@* | node()"/>
>    </xsl:copy>
> </xsl:template>
>
> How do they compare?  I make not of the SAMPLE_001 within the xsl:copy
> there are 2 apply templates, what does each do?

Since <xsl:apply-templates/> is an abbreviation for <xsl:apply-templates  
select="node()"/>, and attributes occur first in document order, the two  
consecutive 'apply-templates' of the first simple do exactly the same as  
the one 'apply-templates' in sample 2.

As for the extra '/' in the second sample, I don't think it is necessary.

> Then, understanding what is happening, is the following sample true
> that this will strip an XML doc of all attribute elements?
>
> SAMPLE_003:
> <xsl:template match="*">
>   <xsl:copy>
>     <xsl:apply-templates select="*"/>
>   </xsl:copy>
> </xsl:template>
Yes indeed, it will strip off attributes BUT also text nodes, comments,  
processing instructions...
(only elements will remain)


> And finally, when I identity transform the following my document i
> quadrupled in size, but I though I was following standard practices
> from previous example.
>
> SAMPLE_004:
> <xsl:template match="*">
>   <xsl:copy>
>     <xsl:apply-templates select="*"/>
>     <xsl:apply-templates />
>   </xsl:copy>
> </xsl:template>


> Ahh, I think I understand what is going on in SAMPLE_004 (please
> confirm) I am basically applying templates for all nodes twice

Not entirely true. You first apply templates to all child element nodes.  
And then you apply all child node (including again the elements)


regards,
--
Joris Gillis (http://www.ticalc.org/cgi-bin/acct-view.cgi?userid=38041)
Gaudiam omnibus traderat W3C, nec vana fides