Re: [xsl] testing for string and number in XSLT 2.0 was Re: [xsl] Test For Numeric Values?

[Dimitre Novatchev <dnovatchev@xxxxxxxxx>]

I repeated your experiment with Saxon 8.4, changing only the variable
definition of $y to:

      <xsl:variable name="y" select="data(example/test/text())"/>

and still got:

       $y variable value: 132131
       Test as string:false
       Test as integer:false

It seems to me that the test as xs:string should be positive, because
the typed value of a text node must be by definition xs:string.

I would appreciate Mike's comments on this.


Cheers,
Dimitre Novatchev



On Apr 9, 2005 7:27 PM, James Fuller <jim.fuller@xxxxxxxxxxxxxx> wrote:
> Michael Kay wrote:
> 
> >>take for example xsl:sort
> >>
> >><xsl:sort select="." data-type="number"/>
> >>
> >>what does 'number' mean here?
> >>
> >>
> >
> >It's retained for backwards compatibility with XSLT 1.0; the "native" way of
> >doing this in 2.0 would be
> >
> ><xsl:sort select="xs:double(.)"/>
> >
> >if they are doubles, or more likely
> >
> ><xsl:sort select="xs:integer(.)"/>
> >
> >if they are integers.
> >
> >
> >
> thx for clarification and generally letting me indulge...its much easier
> to fire off questions whilst reading the specs..
> 
> >>and if we are comfortable with this type of ambiguity of identifying
> >>something purely as a 'number'....then why not have some fairly useful
> >>functions such as is-number() and is-string() that exist in
> >>this no mans land.
> >>
> >>
> >
> >You're welcome to write your own function is-string($x) that has the same
> >effect as the expression ($x instance of xs:string), but the WG wants to
> >avoid bloating the language with redundant ways of doing the same thing.
> >
> >
> agree with avoiding bloat, esp if we are duplicating
> functionality....just a final point....
> 
> using Saxon 8.1b
> 
> EX. XML
> <?xml version="1.0" encoding="UTF-8"?>
> <example>
>    <test>132131</test>
> </example>
> 
> EX. XSLT
> 
> <?xml version="1.0" encoding="UTF-8"?>
> <xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform";
>    version="2.0"
>    xmlns:xs="http://www.w3.org/2001/XMLSchema";>
> 
>    <xsl:output method="xml" indent="yes"/>
> 
>    <xsl:variable name="x" select="example/test" as="xs:integer"/>
>    <xsl:variable name="y" select="example/test"/>
> 
>    <xsl:template match="example">
> 
>        $y variable value: <xsl:value-of select="$y"/>
>        Test as string:<xsl:value-of select="$y instance of xs:string"/>
>        Test as integer:<xsl:value-of select="$y instance of xs:integer"/>
> 
>        I would expect this to be true as I have explicitly cast x as an
> integer
>        $x variable value: <xsl:value-of select="$y"/>
>        Test as an integer: <xsl:value-of select="$x instance of
> xs:integer"/>
> 
>    </xsl:template>
> 
> </xsl:stylesheet>
> 
> I get the following result;
> 
>        y variable value(so we know we are selecting it): 132131
>        Test y as string:false
>        Test y as integer:false
> 
>        I would expect this to be true as I have explicitly cast x as an
> integer
>        x variable value:132131
>        test x as integer: true
> 
> not sure if this is what I would expect normally, the issue is related
> to an element if if has no explicitly declared data-type..
> 
> the above behavior works the same way if I just reference the element
> (example/test) and dont use a variable...though I wanted to highlight
> that there are a few steps where an XSLT Processor may have to *decide*
> what datatype it is.
> 
> it doesnt make much sense to me to *have* to declare something as an
> integer datatype to test if its a value is a number...whats the point?
> 
> I know this is quite a speculative thread and apologies to list for
> going on...
> 
> cheers, Jim Fuller