[MarkLogic Dev General] .NET XCC

[Ron Hitchens]

    It will certainly be longer than a couple of
months.  I can't promise that it will be in any
particular release - that's a policy we have.  I
will say that it's unlikely to happen before 3.2.

On Aug 16, 2006, at 11:46 AM, Wyatt wrote:

> Ron,
>
> Thanks so much for your help.  Makes sense to me...  Any idea when  
> such XDBC protocol improvements might happen?  Is is likely to be a  
> v3.2 enhancement or sometime after that?  I'd hate to write a hack  
> only to have it fixed in a couple months, ya know!
>
> Wyatt
>
> Ron Hitchens wrote:
>> Wyatt,
>>    This should be considered a bug.  The underlying
>> cause is that there is a weakness in the XDBC protocol
>> (XCC speaks the same protocol as XDBC).  There is a
>> presumption in XDBC that node = element and this means
>> that information about the specific node type is not
>> returned by the server (except for text() and binary(),
>> which are handled differently than other nodes).
>>    In XDBC, all nodes (element, attribute, etc) have
>> the same Java type (XDBCSchemaTypes.Node).  In XCC,
>> there are Java types for all the XDM node types.  This
>> means that all node instances will be a concrete
>> type (element(), attribute(), etc) rather than the
>> abstract node() type.
>>    Sadly, resource constraints prevented the needed
>> server-side changes to implement an improved protocol
>> to fully specify concrete node types.  To be consistent
>> with the Java object model XCC assumes nodes are elements,
>> which is the same assumption that exists in XDBC.
>>    You can still obtain the value of attributes by
>> calling AsString() on the XdmItem instance.  The
>> AsString() method should do the right thing for all
>> node types.  If you want to convert an element to
>> a DOM, those methods may throw an exception if the
>> item is not actually an element().
>>    We plan to implement the protocol improvements in
>> a future release.  At that time, the full complement
>> of Java node types will map to the XDM types as expected.
>> On Aug 14, 2006, at 7:54 PM, Wyatt wrote:
>>> Greetings all,
>>>
>>> I've been playing with the .NET version of the XCC package and  
>>> have run into an obstacle I can't seem to surmount.  Maybe a bug  
>>> - more likely my own ignorance...
>>>
>>> I'm trying to write a C# (VS 2005) function that utilizes XCC to  
>>> execute an arbitrary XQuery and return a string of well-formed  
>>> XML.  All is well if the XQuery happens to return a single  
>>> element, but I'm running into trouble post processing it if it is  
>>> not.  It appears that NodeType.ToString() always returns "element()"
>>>
>>> I've been using the XdmItem as so...
>>>
>>> //--
>>> private string ProcessResults(ResultSequence rs) {
>>>     StringBuilder sb = new StringBuilder();
>>>     while (rs.HasNext()) {
>>>         XdmItem xi = rs.Next().Item;
>>>         if (xi.ItemType.Node) {
>>>             sb.AppendLine(((NodeType)xi.ItemType).ToString());
>>>         }
>>>         else if (xi.ItemType.Atomic) {
>>>             sb.AppendLine(((AtomicType)xi.ItemType).ToString());
>>>         }
>>>     }
>>>     return sb.ToString();
>>> }
>>> //--
>>>
>>> The XdmItem.ItemType seems to accurately resolve what is a Node  
>>> and what is an Atomic, and ToString() correctly returns  
>>> "xs:string" for an atomic string (didn't try any others), BUT  
>>> when it is a Node ToString() always returns "element()" even it  
>>> is an attribute node???  I'd have thought it would return  
>>> "attribute()"?
>>>
>>> Any help would be much appreciated.
>>>
>>>
>>> _______________________________________________
>>> General mailing list
>>> General at developer.marklogic.com
>>> http://xqzone.com/mailman/listinfo/general
>> ---
>> Ron Hitchens {ron.hitchens at marklogic.com}  650-655-2351
>> _______________________________________________
>> General mailing list
>> General at developer.marklogic.com
>> http://xqzone.com/mailman/listinfo/general
>
>
> _______________________________________________
> General mailing list
> General at developer.marklogic.com
> http://xqzone.com/mailman/listinfo/general

---
Ron Hitchens {ron.hitchens at marklogic.com}  650-655-2351