Re: [xsl] Need to remove unusual character in source

[Abel Braaksma <abel.online@xxxxxxxxx>]

Mario Madunic wrote:
> I've come across a character in the source that I need to remove.
>
> I'm using Saxon 8, XSLT 2
>
> the character is and its a control character
>
> 0x18 CAN

Perhaps there's another option, if you are really stuck to this 
ill-formed data. You use Saxon, so you might want to use its extensions.

<xsl:function name="try-loading-file">
   <xsl:param name="filename" />
   <xsl:value-of select="document($filename)" />
</xsl:function>

|<xsl:template match="/">
  <root>
   <xsl:copy-of
       select="saxon:try(yourns:try-loading-file('your filename here'),
       saxon:function('yourns:catch-loading-error',1))"/>
   </root>
</xsl:template>|

|<xsl:function name="yourns:catch-loading-error">|

||| <xsl:param name="error-info"/>
   <xsl:sequence select="$error-info"/>
   <!-- DO SOMETHING WITH THE ERROR, LIKE
       CALLING YOUR OWN EXTENSION FUNCTION
       which may take the document, filter it and
       let the application do a retry -->
||</xsl:function>|

||

|This information comes largely from Saxon's documentation: 
http://www.saxonica.com/documentation/extensions/functions/try.html|

||

|Consider this a non-trivial and definitely an unorthodox method. Also, 
your own extension function, written in Java for example, will have to 
create side-effects, which may yield strange results. I don't think this 
is the right path to take, but if you really have to do something with 
this data and if you absolutely have no other option, this might be a 
viable solution.
|

|Cheers,
Abel Braaksma
http://abelleba.metacarpus.com
|

||

|

|