Re: Data types and xsl:sequence

Just specify the types of the parameters -- types help eliminate
exactly such problems:

  <xsl:param name="char" as="xs:string" />

When calling the template pass '0' insted of 0, otherwise you will
(correctly!) get a type error.

So, with minimal changes the transformation becomes something like this:


<xsl:stylesheet version="2.0"
 xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
 xmlns:xs="http://www.w3.org/2001/XMLSchema"
 >
 <xsl:output method="text"/>
  
 <xsl:template match="/">
   Results: '<xsl:text/>
   <xsl:call-template name="pad-number">
     <xsl:with-param name="max" select="5"/>
     <xsl:with-param name="char" select="'0'"/>
   </xsl:call-template>'
 </xsl:template>
 
 <xsl:template name="pad-number">

  <xsl:param name="max" />
  <xsl:param name="char" as="xs:string" />
  
  <xsl:sequence select=
     "string-join((for $i in 1 to xs:integer($max) 
                    return $char
                   ), 
                   ''
                  )"/>
 </xsl:template>
</xsl:stylesheet>

and the result is as expected:

Results: '00000'

The code will be better written as an xsl:function and specifying the
types of all parameters involved, so that type errors will not go
unnoticed:

<xsl:stylesheet version="2.0"
 xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
 xmlns:xs="http://www.w3.org/2001/XMLSchema"
 xmlns:t="http://myURL/test"
 >
 <xsl:output method="text"/>
  
 <xsl:template match="/">
   Results: '<xsl:value-of select="t:pad(5, '0')"/>'
 </xsl:template>
 
 <xsl:function name="t:pad">
  <xsl:param name="max" as="xs:integer" />
  <xsl:param name="char" as="xs:string" />
  
  <xsl:sequence select=
     "string-join((for $i in 1 to $max 
                           return $char
                          ), 
                         ''
                        )"/>
 </xsl:function>
</xsl:stylesheet>

Note how the code is actually simplified:

  1. No need to have a special template for pad-number and pad-character

  2. No need to construct an xs:integer from $max

  3. Benefits from type checking so that if inappropriately-typed
arguments are passed this will not go unnoticed and will be
immediately raised as error

  4. Referencing the function is much more compact than the 4 lines
needed for xsl:call-template.

  5. More importantly, the function can be used immediately in any
XPath expression at any place where the syntax rules allow this.


Cheers,
Dimitre Novatchev.
 


On Mon, 24 Jan 2005 16:17:41 -0500, Jim Neff <jneff@xxxxxxxxxxxxxxx> wrote:
> Greetings,
> 
> I'm having trouble understanding data types and using the function
> xsl:sequence.
> 
> I have one common template that I call to create padding in my output
> document:
> 
> <xsl:template name="pad-number">
> 
>   <xsl:param name="max" />
>   <xsl:param name="char" />
> 
>   <xsl:sequence select="string-join((for $i in 1 to xs:integer($max) return
> $char), '')"/>
> 
> </xsl:template>
> 
> So if I am padding for a text field, I pass in &#x20; for a space and if
> it's a numeric field I use '0'.
> 
> My problem is (and I know this sounds silly) whenever I change from a text
> to a numeric field there is an extra space inserted into the xsl:sequence
> output.  This is a space and not the character I am passing into this
> template.  So, if I use a tilda '~' instead of '&#x20;' I see the
> appropriate number of tildas plus a space (hex code 20) in my output.
> 
> I do not have this problem when I am not changing from numeric to text, or
> if I have consecutive numerics or consecutive text fields.
> 
> Perhaps I am not using the $max variable correctly?  I found the only way to
> get it to work is when I specify xs:integer around it because the portion of
> code that calls this template is performing a calculation and I think the
> processor (Saxon 8.something) is creating a mandotory xs:double whenever a
> calculation is performed.
> 
> This is probably a logic problem and has nothing to do with XSLT syntax but
> I just thought I'd throw this out here and maybe one of the gurus on this
> list could point me in the right direction.
> 
> Thanks,
> Jim Neff