Hi,
It seems that the current error handling resets the FlowFile's original content..
So:
I fixed it by throwing exception so the original FlowFile won't be rewritten and then transfer it to failure.
I added a capability to preserve the field names
I added a capability to process messages separated by a demarcator(like when consuming using the ConsumeKafka processor)
I added a new processor that converts the Protobuf messages into Avro messages so they would be easier to handle with NiFi(rather than JSON that doesn't preserve the original schema)
I changed the processors' names to EncodeProtobuf and DecodeProtobuf to match NiFi conventions
Hi, It seems that the current error handling resets the FlowFile's original content.. So:
Hope it's good enough!