We get one large EDI file with thousands of ISAs in it.
Our split option in EDITPA is set to “Transaction”
When we submit this file, the recognize process takes a long time (up to an hour). I am guessing it iterates through ISAs one by one, splits them in groups and transaction and persists them.
If I set the split option to “Interchange” then it runs much faster (5-6 mins).
So can we
Configure the recognize to process the ISAs concurrently? I didn’t find any settings for this.
OR
Change the split option in EDITPA to Interchange and then manually split an Interchange in a processing service. Has anyone done this before? What would it take?
" Change the split option in EDITPA to Interchange and then manually split an Interchange in a processing service. Has anyone done this before? What would it take? "
–Yes if you are processing it Envelop level (Interchange) you should see a var called editn_env (in the pipeline) and downstream use envelopeProcess service to get EDIValues and then you will need to loop over IS/GS to get the ST level data via getEDIString and proceed further…
Since you mentioned Large EDI files…have you also configured large file handling for TN and EDIBizDocThreshold to retrieve from the TSpace for processing IB EDI data?
I have not. The only thing I am not sure is how will I submit the individual transactions to TN and make them appear as say “X12 4010 997”.
For example, for an EDI with one ISA, GS and ST, TN recognize does the split and creates X12 Envelope, X12 Group and (say) X12 4010 997 document types.
If I manually parse out the GS and the STs and create individual envelopes for each and submit them to TN, there is no way to specify that the document type being submitted is an “X12 4010 997” and not an “X12 Envelope”. Specifying a document type is only allowed for flat files.
Even when trying to process “X12 4010 997” to TN you still need to add Envelopes before routing to TN recognition and I know you want see the visibility but it doubles your down stream processing isn’t? I would rather process each ST level mapping or let EDITPA split the Transaction level which may not be feasible to you from the above scenario…