Kevin Hill created NIFI-6876:
--------------------------------
Summary: Sub-record parsing
Key: NIFI-6876
URL: https://issues.apache.org/jira/browse/NIFI-6876
Project: Apache NiFi
Issue Type: Improvement
Reporter: Kevin Hill
After using Nifi for awhile it seems like the main boundary that forces our
team to create a custom processor is when some field in a data source is
actually it's own encoded record or array.
A common example might be a json object stored in a Postgres SQL database. We
can get the row as a record, but that field will be encoded as a JSON string,
which means we are unable to reference the underlying values of that sub-record
in any easy way in Nifi. So, to really deal with any of those values, or
potentially create one-to-many mappings similar to ForkRecord, we have to go
straight to a custom script or processor.
What I imagine is a processor that allows dynamic properties with a RecordPath
as a key, and a RecordReader as a value, but anything that achieves the same
end would work.
I'm 99% sure this feature doesn't exist in Nifi, but if I'm mistaken I would be
very happy to hear that. I think it would vastly extend the capabilities of the
types of records nifi could process without relying on writing custom script
processors.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
