Github user zenfenan commented on the issue:
https://github.com/apache/nifi/pull/2751
Actually after giving it some serious thought, the attribute approach does
make more sense. Easy from the flow development perspective, and even when
using ExecuteScript or a custom processor to parse and process content to
create appropriate tags, the developer can still add them to FlowFile
attributes.
So Iâll update this one to scrap the JSON approach and instead take a
comma
separated list of tag prefixes which will be taken and scanned against the
incoming FlowFilesâ attributes to generate the tags. Makes sense?
---
