when the duplicate rosettanet transaction is coming on to the process model, how does this will be detected.
Are you using RNIF in IS/TN/modeler or IS/TN environment?
Duplicate transactions can be avoided basing on the TN Docid or Converstationid or co-relationid (Modeler) which maintains uniqueness technically.
If this is not useful, please elaborate your existing architecture, so we can help.
I am using RNIF in IS/TN/modeler. I know to handle duplicates in TN environment using processing rules. I am trying to findout how to eliminate the duplicate transactions on modeler. Since the process model i am going to generate is going to subscribe the TN doctype, we are not using processing rules. So i am just trying to findout, what happens when the duplicate document comes on to TN and process model subscribes the dupliate document again. Now how will it find the duplicate transaction.
I will appreciate your help in this regard.
Thank you rmg,
you can make use of co-relationid’s (same like TN conversationid) which keeps track of all incoming documents in the process model and this can eliminate the duplicates.
Check the modeler documentation about managing of corelationid’s.
You mean to say, use correlation services? or whether the PRT can handle it by itself?
you have to use correlation services,and depend on this only PRT can handle it and not by default.
One solution may be to use modeler to create a process step that can only be executed once. The duplicate document would hit this step and be sent to the retries exceeded path. The retries exceeded path will lead to a process step that could “suppress” the document (using join conditions on the step) or rejoin the already running process (depends on your exact implementation).
Just some thoughts.
I’m really confused by this thread. What’s the definition of “duplicates”?!? There are only two IDs that could be checked, AFAIK – ConvID, and DocID. Or are we talking about having different TN DocTypes with the same DocID being duplicates?
ConvID - If you’re using wM-supplied model (One action Responder, for example), receiving duplicate docs with same ConvID is already handled for you. Your service (mapping or anything after it) won’t even see the doc to be able (or need to) check for duplicates.
DocID - I thought duplicate DocIDs are checked as part of the document recognition process (before processing rule or beginning of model process? Actually for inbound RN docs, Processing Rules are skipped by RN’s receive, so that won’t occur for sure.
Or am I so off base here because Raj is using his own models to process inbound RN docs?
Also, TN 6 (at least in 6.1, I haven’t bothered checking 6.01) have the ability to invoke a custom service to check uniqness (in DocType definition). This seems to me to be much preferable than actually trying to check duplicates w/in models.