Github user timhallinflux commented on the issue:
https://github.com/apache/nifi/pull/2743
PutInfluxDB was created as single purpose tool -- accepting line protocol
only. It does not read CSVs, AVROs, JSONs out of the box. If developer wants to
use it with NiFi, s/he has to extend it and write their own parsers. That is
ok...and it does support a direct integration.
However, PutInfluxDBRecord addresses the problem of reading the data. It
simplifies data parsing and handling using the concept of Records (
https://blogs.apache.org/nifi/entry/record-oriented-data-with-nifi )
It more naturally works with Nifi objects and the fields, tags are
configurable from within the Nifi IDE...which delivers a much more integrated
experience.
For example: Reading Twitter with PutInfluxDB is not possible without
coding/external configuration. On the contrary, reading Twitter JSON via
PutInfluxDBRecord is no-dev effort and leverages the tools within NiFi itself.
Still, the two classes can coexist depending on the type of work that needs
to be done. I'm in favor of moving ahead with the PutInfluxDBRecord
---
