Subject:BizTalk schema problem Author:Kurt Wald Date:24 Mar 2007 03:32 PM Originally Posted: 24 Mar 2007 03:26 PM
Apparently due to xerces schema validation issues several valid xml schemas cannot be loaded into the stylus studio mapper as source or target document(e.g. see example).
Would it be possible to use a less problematic engine for mapper construction or are there any workarounds except major manual schema modifications, which are not really an option in a production environment?
Subject:BizTalk schema problem Author:(Deleted User) Date:26 Mar 2007 11:12 AM
Hi Kurt,
from what I can see the problem is limited to one known bug in Xerces (the usage of namespace prefixes created in the middle of the document); I don't know how BizTalk generates its schemas, but could you check if it has a way to add a global binding for the namespace http://DELFOR01.V3 in the root element? Maybe, if such a binding is present, all the q1, q2, qN prefixes would not be generated and Xerces could process it.
Subject:BizTalk schema problem Author:Kurt Wald Date:26 Mar 2007 02:04 PM
Hi Alberto,
I have posted a schema about a month ago which ran into the Xerces default value 0 in enumeration bug, you have created a workaround for the Xerces large finite loop bug, just to name a few. According to my understanding we only talk about the parser used to build the document trees for the XSLT mapper here. Document validation does not seem to be the major requirement in this regard. So a more "liberal" way of building the trees might be helpful. I have talked to a number of developers with no or little knowledge of XML schema, who tend to question the mapping product, if they can't get the mapper up and running with an XML schema, which is valid and produces no error messages with Microsoft, Oracle and other validation engines.
Subject:BizTalk schema problem Author:(Deleted User) Date:27 Mar 2007 08:19 AM Originally Posted: 27 Mar 2007 08:17 AM
Hi Kurt,
I understand your feeling, but unfortunately XMLSchema is so complex that you cannot extract the structure of the document unless you use a real XMLSchema parser; anyhow, we have fixed the problem you reported last month, and we'll try to fix this one too.