In Boomi, I have a process that:
a) queries a SQL server database table for 1 or multiple records (at the moment, the batch setting on the database operation is set to 0)
b) check each of the records key value against the vendors API (Successfactors)
-if record exists, using a flow control, batch up to 10 records, and then go ahead and UPDATE the records into the vendor system using API
-if record does not exist, write error to SQL Server database table, stop and continue
Now what is happening is:
In the case I have only one record to UPSERT, and that record is invalid, the process gracefully fails the check, writes the error to the database, and stops as expected.
In the case I have two records, one valid and one invalid, both records pass the check for validity and both are reported as successfully Updated into the system - which is impossible since the invalid record does not exist in the API!
Now, if I put a flow control before b) above, and run each document individually, then the valid record is successful in updating, and the invalid record is unsuccessful - correct behavior. However, for 100's of records, the performance degrades sharply and is unacceptable (I am updating from a local atom to an external vendor API)
Any help is appreciated