Global elements are called open content properties in SDO. The 2.0.1. spec 
says that a property is from open content if it appears in 
getInstanceProperties() but not in getType.getProperties(). One way to 
check this: !d.getType.getProperties().contains(p). Another more efficient 
way is !p.getContainingType().isInstance(d). In SDO 2.1, there is going to 
be a new method added, to make this even easier: Property.isOpenContent().

Frank.

"Yang ZHONG" <[EMAIL PROTECTED]> wrote on 09/23/2006 03:29:49 
PM:

> One of my assumptions was wrong: I thought XSDHelper#getGlobalProperty 
could
> be used to help to distinguish local element from global element without
> NameSpace.
> 
> I don't think so now since a no-NameSpace global element existence does
> *not* prove my Property is an element ref to that.
> 
> I can't find a way in current API to distinguish local element from 
global
> element without NameSpace.
> It's blocking one of my work (ChangeSummary StAX writer) unless I use 
EMF.
> 
> Can any one help please?
> 
> 
> On 9/23/06, Yang ZHONG <[EMAIL PROTECTED]> wrote:
> 
> > Local XML element doesn't have a NameSpace,
> > neither does global XML element without NameSpace.
> > It'll be nice for XSDHelper#getNamespaceURI to return "" for local XML
> > element (since "" is an *invalid* NameSpace)
> > and to return null for global element without NameSpace (since null is
> > also used to query Types and Global Properties without NameSpace).
> > Is it possible for the Java/C++/PHP specs to clarify that please?
> >
> > Otherwise, is it possible for the Java/C++/PHP specs to claify 
returning
> > null or "" for both cases
> > and to tip users to make another round trip such as
> > XSDHelper#getGlobalProperty if necessary please?
> >
> > I'll blame myself if this note has ever any influence to lead either
> > of the Java/C++/PHP specs to ending up with clarifying "either 
behavior is
> > fine with each implementation".
> >
> > --
> >
> > Yang ZHONG
> >
> 
> 
> 
> -- 
> 
> Yang ZHONG


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to