Conversation ID for EDI docs not being set as expected

I want to link EDI documents in TN using the conversation ID. My problem is that TN is creating new conversation IDs for each transaction-level document, even when the instance IDs are the same. For example, an inbound 850 is received and gets a conversation ID of senderID-receiverID-instanceID. An outbound 870 with the same instanceID (PO number) is later received and gets a new conversation ID, even though the docs say that for the second document it should try to use an existing ID by flipping the sender and receiver to search for it.

Where is TN looking for the existing ID – documents or process models? I have not created a business process in Modeler for this, and would prefer not to if possible. Thanks,

Tim

I’m reposting this as I never got the e-mail with the original post myself. I apologize if you already received this and are simply ignoring my plight :wink:

I want to link EDI documents in TN using the conversation ID. My problem is that TN is creating new conversation IDs for each transaction-level document, even when the instance IDs are the same. For example, an inbound 850 is received and gets a conversation ID of senderID-receiverID-instanceID. An outbound 870 with the same instanceID (PO number) is later received and gets a new conversation ID, even though the docs say that for the second document it should try to use an existing ID by flipping the sender and receiver to search for it.

Where is TN looking for the existing ID – documents or process models? I have not created a business process in Modeler for this, and would prefer not to if possible. Thanks,

Tim

Can anyone out there using conversation IDs tell me why I’m getting a new ID even when the instance IDs are the same and TN should be finding the existing one? Thanks,

Tim

Here’s the resolution, in case anyone finds themselves in the same situation later. According to wM support, the conversation ID is only used when PRT is involved. To use the conversation ID without PRT I did the following: 1) added an index to the Bizdoc table in the TN database for ConversationID; 2) wrote a JDBC adapter service to query the Bizdoc table with SenderID, ReceiverID, DocTypeID and ConversationID as inputs and DocIDs as output; and 3) flipped the sender and receiver on the conversation ID, stored this in a variable and used this as input to the service. The results give me any documents between the same partners with the same instance ID, in this case PO number.

Tim