Github user JohannesDaniel commented on the issue:
To be aligned with processor GetSolr I implemented two options for the data
format of Solr results: Solr XML and record functions. However, facets and
stats are written to flowfiles in JSON (which has the same structure like the
Solr-JSON). I did not implement record management for these two components to
keep the complexity of the processor at a reasonably level. I chose JSON as it
is probably the best integrated format in NiFi.