Hi Neil,

You generally find that, ascross transformation engines, explicitly
'casting' types yields between none and some performance improvement ;).
With MS's transformation engines you tend to see the best performance hike
for these subtle changes - my guess would be the way they internally deduce
data types before deciding whether a conversion is needed.

Of course, there are times in XSLT when explicitly casting is an absolute
must in order to get the desired result - the one example that always
springs to mind is index predicates, e.g. the possible difference in results
between...

<xsl:variable name="index" select="1"/>
<xsl:value-of select="count(foo[$index])"/>

and...

<xsl:variable name="index">1</xsl:variable>
<xsl:value-of select="count(foo[$index])"/>

Cheers
Marrow


"Neil Smith [MVP Digital Media]" <neil@n...> wrote in message
news:nootj01v43b99su91s2dfttvpgo1ldhdit@4......
> Thats a curious result, Marrow - are you saying that explicity telling
> the MSXML parser to 'type' the data gives a performance improvement
> over letting it implicitly determine the data type, in this case ?
>
> I wonder, is this a general method for improving  XSLT parse
> performance, or is it specific to MSXML ?
>
> Cheers - Neil
>
> On Wed, 8 Sep 2004 10:38:26 +0100, "Marrow"
> <m--a-r-r-o-w@m-a-r-r-o-w--s-o-f-t.com> wrote:
>
> >Hi Darren,
> >
> >I think you might be surprised at the performance 'difference'...
> >
> >If you tweak the code I gave by replacing...
> >   <xsl:param name="n" select="."/>
> >with...
> >   <xsl:param name="n" select="string(.)"/>
> >and replace...
> >      <xsl:variable name="nv" select="substring-before($n,'E')"/>
> >with...
> >      <xsl:variable name="nv" select="number(substring-before($n,'E'))"/>
> >
> >that should speed it up.
> >
> >Compared with using msxsl:number() on MSXML 4.0 over a sample of your
data
> >repeated 20,000 times - I am seeing almost identical times (averaged out
on
> >one of my machines at:- template method = 173.8ms; msxsl:number() method
=
> >173.2ms).
> >
> >Of course, be aware that msxsl:number() is not consistently supported
even
> >as an MS extension (it isn't supported in MSXML 3.0).
> >
> >Cheers
> >Marrow
> >
> >
> >"darren@w..." <darrenworkcom@d...> wrote in
message
> >news:6EED1412-0908-4AAE-A9A6-8943ECAFF88D@m......
> >> Thanks for the reply,  I have actually gone with Joe's solution above,
as
> >our
> >> datafiles can be quite large, and  predominantly numbers.  I there
suspect
> >> the petrformance of further template calls (especially the number that
> >would
> >> be required) if a function is available.
> >>
> >> Thanks for the post,
> >>
> >> Hugh
> >>
> >> "Marrow" wrote:
> >>
> >> > Hi Darren,
> >> >
> >> > You could write a template to handle exponential notation numbers,
> >something
> >> > like...
> >> >
> >> > <?xml version="1.0" encoding="UTF-8"?>
> >> > <xsl:stylesheet version="1.0"
> >> > xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
> >> > xmlns:fo="http://www.w3.org/1999/XSL/Format">
> >> >
> >> > <xsl:template match="stuff">
> >> >   <xsl:text>Value: </xsl:text>
> >> >   <xsl:call-template name="exnumber"/>
> >> >   <xsl:text>
> >> </xsl:text>
> >> > </xsl:template>
> >> >
> >> > <xsl:template name="exnumber">
> >> >   <xsl:param name="n" select="."/>
> >> >   <xsl:choose>
> >> >     <xsl:when test="contains($n,'E')">
> >> >       <xsl:variable name="nv" select="substring-before($n,'E')"/>
> >> >       <xsl:variable name="nexp" select="substring-after($n,'E')"/>
> >> >       <xsl:variable name="exp"
> >> >
>
>select="substring('1000000000000000000000000000000',1,translate($nexp,'-','
'
> >> > )+1)"/>
> >> >       <xsl:choose>
> >> >         <xsl:when test="$nexp >= 0">
> >> >           <xsl:value-of select="$nv * $exp"/>
> >> >         </xsl:when>
> >> >         <xsl:otherwise>
> >> >           <xsl:value-of select="$nv div $exp"/>
> >> >         </xsl:otherwise>
> >> >       </xsl:choose>
> >> >     </xsl:when>
> >> >     <xsl:otherwise>
> >> >       <xsl:value-of select="$n"/>
> >> >     </xsl:otherwise>
> >> >   </xsl:choose>
> >> > </xsl:template>
> >> > </xsl:stylesheet>
> >> >
> >> >
> >> > HTH
> >> > Marrow
> >> > http://www.marrowsoft.com - home of Xselerator (XSLT IDE and
debugger)
> >> > http://www.topxml.com/Xselerator
> >> >
> >> >
> >> > "darren@w..." <darren@w...@discussions.microsoft.com> wrote
in
> >> > message news:B41E4F0B-C97F-452F-8220-1326DE921065@m......
> >> > > I have an xml file:
> >> > > <?xml version="1.0" encoding="UTF-8"?>
> >> > > <myspace>
> >> > > <stuff>5.6</stuff>
> >> > > <stuff>1.27E-2</stuff>
> >> > > </myspace>
> >> > >
> >> > > I have a stylesheet:
> >> > > <?xml version="1.0" encoding="UTF-8"?>
> >> > > <xsl:stylesheet version="1.0"
> >> > > xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
> >> > > xmlns:fo="http://www.w3.org/1999/XSL/Format">
> >> > >
> >> > > <xsl:template match=".//stuff">
> >> > > Value: <xsl:value-of select="number(.)"/>
> >> > > </xsl:template>
> >> > >
> >> > > </xsl:stylesheet>
> >> > >
> >> > >
> >> > > I expect the results:
> >> > > <?xml version="1.0" encoding="UTF-8"?>
> >> > > Value: 5.6
> >> > > Value: 0.0127
> >> > >
> >> > > The results I get:
> >> > > <?xml version="1.0" encoding="UTF-16"?>
> >> > > Value: 5.6
> >> > > Value: NaN
> >> > >
> >> > > The parser doesn't seem to support  exponential notation for
numbers!
> >> > > Is this true, or am I doing something wrong?
> >> > > Is there a work around, as I have a lot of  data from excel
> >spreadsheets
> >> > > that use this notation, and my style sheets are not working well
with
> >> > it!!!
> >> > >
> >> > >
> >> > > Hugh
> >> > >
> >> >
> >> >
> >> >
> >
>