[
https://issues.apache.org/jira/browse/NIFI-16399?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18121193#comment-18121193
]
David Handermann commented on NIFI-16399:
-----------------------------------------
[~mike.thomsen] Thanks for continuing the discussion here.
I understand the goal of providing a way to communicate FlowFile attributes
back to NiFi, but I don't agree with using temporary files for the
implementation. Even optional features need to be evaluated for complexity and
maintainability, so I'm not in favor of going forward in this direction as it
stands.
One possibility is to have a different contract of the standard output from the
script. That would require some level of design consideration, but it could be
something simple, perhaps analogous to HTTP with a header section and a body
section. At that point however, it may also be better as a separate Processor,
as opposed to an optional setting in this one.
The main concern here is introducing complexity for an optional feature, which
creates problems for future maintenance. Having a defined way for a generic
script to return both FlowFile attributes and content makes sense, but I think
it needs some additional design consideration before just adding it in this way.
> Enable attribute updates from ExecuteStreamCommand
> --------------------------------------------------
>
> Key: NIFI-16399
> URL: https://issues.apache.org/jira/browse/NIFI-16399
> Project: Apache NiFi
> Issue Type: Improvement
> Reporter: Mike Thomsen
> Assignee: Mike Thomsen
> Priority: Major
> Time Spent: 1h 40m
> Remaining Estimate: 0h
>
> This change will create an optional mode where ExecuteStreamCommand will
> write a temp file to the file system that can be used by the executed command
> to communicate updates to the flowfile attributes. It will be a flat json
> document.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
