CDATA section

[CodeDigger]

I have a <Any> element in my Schema to catch any type of content -

<xs:element minOccurs="0" maxOccurs="unbounded" name="MixedDescription">
<xs:complexType mixed="true">
<xs:sequence>
<xs:any processContents="skip" />
</xs:sequence>
</xs:complexType>
</xs:element>

But it does not catch entities like this &ndash; -

<MixedDescription>....some content &ndash; ... content continues</MixedDescription>

XmlSpy, Stylus Studio and Visual Studio all point out that this is an undeclared charater entity.

I came across this idea to implement entites as elments -

http://www.topxml.com/code/default.asp?p=3&id=v20010829094626

but there are a LOT of characters to replace which are possible Xml Character entities, so this is not a practical solution.

What wrapping values with CDATA clause works, but how to declare that in a Biztalk schema? How to instruct Biztalk that one node is a CDATA node and ignore its contents?

 

[CodeDigger]

OK I figured this out via this hint

http://searchwebservices.techtarget.com/tip/1,289483,sid26_gci879720,00.html

Use an XSLT library called xmlchar

http://www.xml.com/pub/a/2003/01/02/xmlchar.html

and in the instance add this line (via pre-processing) (where xmlchar is the href to the library URI)


Code:
<!DOCTYPE ONIXMessage[<!ENTITY % html.4.entities SYSTEM "xmlchar/html4-all.ent">%html.4.entities;]>
<ONIXMessage xmlns:xsi="

http://www.w3.org/2001/XMLSchema-instance"

xmlns:ch="

http://xmlchar.sf.net/ns#"

...(other declarations) >


Don't forget to add the xmlns:ch declaration and make sure the DOCTYPE root node is your root node.