06-20-05 12:47 PM
After some testing, it looks like the Orchestration caches the message or
something similar. I found this out by cloning the XmlDocument (using
CloneNode) and then assigning it to the SQL Request message parameters.
"Re-setting" the parameters so to speak. Doing this re-set the message is
then sent in its full to the database.
/Chris
"Christoffer" <christoffer@nospam.com> wrote in message
news:%23niOPzXdFHA.384@TK2MSFTNGP10.phx.gbl...
> Hello, I have quite a strange behaviour in my Orchestration right now and
> I hope someone could shed some light on this issue.
>
> To simplify; I receive a message and send it to a SQL Server. I've added
> the multi-part message/schema using "Add generated items..." and I'm using
> a Map to create the SQL Request message. Using this solution its all fine,
> everything works ok.
>
> Now, I do the following:
> 1. I remove one of the three fields from the Map, thus only two fields are
> mapped from the incoming message to the outgoing SQL Request message.
> 2. I'll create a XmlDocument from the SQL Request message, and in this
> XmlDocument I append the field which was removed from the map.
>
> Thus, instead of using a Map to set the field in the SQL Request message,
> I use XmlDocument to append a XmlAttribute. I send it, but the SQL Server
> complains that the field has not been set.
>
> Here is what I found out so far:
> 1. The SQL Server does not receive the third field (found out by using a
> trace).
> 2. The SQL Request message does contain all fields (found out by using
> log4net).
>
> I had to re-check number 2 again, basically because I believe this is
> where the error is. Just before calling the Send shape with the SQL
> Request message, I send it to log4net. The multi-part message contains
> only one part which is the SQL Schema. This one part is exactly the same
> no matter if I use a Map or a XmlAttribute to create it.
>
> Somehow, I believe my changes to the SQL Request message is lost when
> using XmlAttribute but I can't figure out how or why. Is there some hidden
> information in the multi-part message or a cache of some sort?
>
> Sorry for the long post, but I'm quite confused here...
>
> Regards,
> Chris
>
>
[ Post a follow-up to this message ]
|