Forum Replies Created
-
Posts
-
What adapter are you using to send out the message? the FILE adapter?
Let me see if I understand what you’re saying: You say that if you remove all the ZIP’ing code and just trace the message as it gets to your pipeline component, you see something that looks like it’s UTF-16 encoded, but when the message finally comes out of the adapter, it looks like if it was UTF-8 encoded, right?
If so, does this happen too if you completely remove your custom pipeline component from the equation?
One final thing: How is your custom pipeline configured? (i.e. what components and at which stage?)
I think so, because this is the begine of my input file in hex:
FFFE 3C00 3F00 7800 6D00 6C00 2000 7600 6500 7200 7300 6900 6F00 6E00 3D00 2200 3100 2E00 30000 2200
and so on.
The Charset property of the BodyPart is set with \”null\”, so I think that is cleared.
but have you ever tried this?
I set a Return pInMsg
to the end of the code and repace the Return OutMsg with this.
So i told the component to give the stream, it gets from the pipeline, back to the pipeline without any processing.
And than I start the pipeline.exe with a Zip file in place of the XML file as input.
In this constallation I thought that the output file must be the same as the input file, because there is no processing on the file.
But there was the same effect like I create my zip files from the xml file.There isn’s such thing as a hidden assembler (besides, assembly happens before encoding in send pipelines).
However, it is possible that the adapter might interpret that somehow. At least, you might want to try clearing out the Charset property of the BodyPart, that might help somewhat.
That said, how do you know you’re giving the pipeline an UTF-16 encoded file originaly?
I tried you suggestion.
I delete the pContext.ResourceTracker.AddResource(OutputStream); entry but the result was the same.
The file the pipeline leave is not equal than the file I give back to the pipeline.
I think there must be a hidde coder/assembler in the end of the pipeline.
I Think so because when i give my input-file, a UTF16 coded XML file, back to the pipeline, the output of the pipeline is a UTF8 coded XML file.
And there is no other coder/assembler component in the pipeline as my zip component.So how can I tell the pipeline that the stream i give back to the pipeline is a binary file and not a text or XML file?
I’m kinda curious… has anyone tried defining the envelope schema twice? one marked as an envelope schema and one as a regular document schema; and just doing the map against the document-one and doing the debatching against the envelope one. A pita, but it might work….
This is usually a problem with something else – not the Xml Documents.
Check your .net code. I’d pass Xml Documents in and get Xml Documents (or strings) out. Then, inside your code get the outer Xml.
Also, make sure you are not passing any complex types with Xml Documents. That will give that error as well.
Also, make sure you set your Xml document back to a message of type Xml Document before you send it out.
Hope this helps.
You posted in BT2004 forum. BT2004 does not have a built-in Oracle adapter but 2006 does. If you need an Oracle adapter for BT2004, check http://TopXml.com or other vendors.
If you are asking aobut the Oracle Adatper in BT2006 – please specify, but I don’t know anything about it. Hopefully someone else can answer.
Neal
Thanks a lot, that got me going. What’s confusing is there is the Covast Trading Partner Management and the BizTalk Parties which have different values in them. Also the EDI Adapter Send Properties I set the EDI Address to EDI://:: I don’t know if this is necessary or right but it works.
Versioning is always had in BizTalk. In fact, I’ve never been in a situation that we had to version.
I can’t think of a good way off the top of my head to accomplish your solution. I think it’ll be hard to have a single server to support both A1 and A2 without making code changes on your client side.
If you wanted your service to send to A1 and then switch to A2, I think that would be more doable by modifying the way the web service published the message to the message box. See my sample on Submitting Xml Data via SOAP.
Hope this helps.
I have receive location that generates this stored:
select fldid,flddate,fldflag from tblUpdateADLog where fldflag = 0 for xml auto,xmldata
update tblUpdateADLog set fldflag = 1
insert into tblUpdateADLog (fldflag) values (0)I distinguish the fldid field.
Every 15 sec. interval, BT is executes the stored and new row inserted.
The only problem is that I can’t get the distinguish fldid value.msgIn.tblUpdateADLog.fldid
\”Im getting: object reference not set to an instance of an object\”
Im using xmlrecieve as pipeline type.Any idea?
That was my first thought too. Problem is, I’m not getting anything returned to extract. I need to match on last_name so if the incoming data is ‘SALAZAR’ it does not equal ‘Salazar’ therfore no value from the db, Oracle BTW. What I need is to do simulate \”select (*) from customer where upper(last_name) = ‘SALAZAR’\”.
Thanks
One very useful case to use a ServicedComponent from an orchestration is to ensure transactional calls. If the component is marked as supporting transactions, and it is called from an atomic transaction scope in the orchestration, biztalk will make the scope into a DTC transaction and run the .NET component inside that same transaction.
Maybe that is what they are doing. Or maybe they depend on a different COM+ service. Maybe they are doing it just so they can execute it inside a Server COM+ application running under a particular identity (i.e. for security reasons) and so on.
As Stephen says, you’d need to ask whoever created it; I just wanted to point out that there are perfectly good reasons for using EnterpriseServices.
here is an example by Stephen Kaufman the illustrates conditional mapping. this should give you a general idea that should steer you in the right direction. of course, you would have logic instead of a simple \”exists\” or \”= ‘United States’\” but youll get the idea.
If you still cant get the solution working, send me a pair of files and I’ll whip you something up.
Welcome to our world lucky027 😀
I take that back, the database lookup functoid does nothing more than return a record set from your table based on a lookup value (primary key for example).
I figured there would be some way of manipulating the output through the value extractor functoid in the second half of the operation but i was wrong.
you must take the output from the value extractor and feed it to the uppercase functoid.ok that seems pretty standard.
go ahead and type in EDI:\\\\:: ( a catch all URI for edi) inside your recieve pipeline. This will correspond to an alias within your party and basically acts like a * ( in file masks for example ) for any other party alias within that party. ** this is what covast means when they say that the sender URI is the only required one. **note**** depending on the build you are using 1.03, 1.07 or 1.20 this may or may not work. if it doesnt work, then delete that all together from your pipeline config and just leave it blank… it will basically substitute whichever party you have configured in biztalk that comes in that file.
from the descriptions that you have given (one working and the other not) i would say that you have that particular party in the pipeline config, hence, one works and the other doesnt. but everything else youve provided, included the samples, seem to be in order.
