NIFI-6970 add DistributeRecord processor

[IlyaKovalev]

Thank you for submitting a contribution to Apache NiFi.

Please provide a short description of the PR here:

add DistributeRecord processor for distribute data over user specified relationships by distribution key/keys.

In order to streamline the review of the contribution we ask you
to ensure the following steps have been taken:

For all changes:

• Is there a JIRA ticket associated with this PR? Is it referenced
  in the commit message?

• Does your PR title start with NIFI-XXXX where XXXX is the JIRA number you are trying to resolve? Pay particular attention to the hyphen "-" character.

• Has your PR been rebased against the latest commit within the target branch (typically master)?

• Is your initial contribution a single, squashed commit? Additional commits in response to PR reviewer feedback should be made on this branch and pushed to allow change tracking. Do not squash or use --force when pushing to allow for clean monitoring of changes.

For code changes:

• Have you ensured that the full suite of tests is executed via mvn -Pcontrib-check clean install at the root nifi folder?
• Have you written or updated unit tests to verify your changes?
• Have you verified that the full build is successful on both JDK 8 and JDK 11?
• If adding new dependencies to the code, are these dependencies licensed in a way that is compatible for inclusion under ASF 2.0?
• If applicable, have you updated the LICENSE file, including the main LICENSE file under nifi-assembly?
• If applicable, have you updated the NOTICE file, including the main NOTICE file found under nifi-assembly?
• If adding new Properties, have you added .displayName in addition to .name (programmatic access) for each of the new properties?

For documentation related changes:

• Have you ensured that format looks appropriate for the output in which it is rendered?

Note:

Please ensure that once the PR is submitted, you check travis-ci for build issues and submit an update to your PR as soon as possible.

[shayneburgess]

@IlyaKovalev I had a couple of quick questions looking at this:

1. Are you going to add this to the manifest for the standard processor NAR?
2. Distributing based on a field is really interesting and useful but the DistributeLoad also has the ability to round robin distribute. Did you think about doing that here? Potentially make the Keys field optional and do round-robin if not specified. I'm thinking of a case where I don't want to distribute based on the actual data.

[IlyaKovalev]

1 Fixed, also remove ORIGINAL relationship (it's excessively i think)
2 Yes, i really thought about it. In case merge logic of DistributeRecord with DistributeLoad we should be add reader, writer, keys, hash_function, strategy fields but logic for understanding how processor will process input is growing i suggest it's too ambigious.
So look at this like:
DistributeLoad - works on flowfile level without processing content i.e distribute flowfiles.
DistributeRecord - works on content level i.e distribute content
(So maybe we need rename DistributeLoad to DistributeFlowFile just because DistributeLoad is
very general definition)
I think it's much easier for understanding processor logic(how it would works) and much simplest for user to appreciate init process.

[joewitt]

Hello. This is an interesting PR/good idea. I wonder though if this should just be a partitioning strategy in PartitionRecord. If we keep it separate like this we might want to go with a more specific name than DistributeRecord as we might want different distribution options and this one seems to be pretty specific to weighted distribution using one or more hashing functions. Maybe then the name should be 'WeightedRecordDistribution'

In any case how about squashing the commits and rebasing to master and pushing the forced PR. This will let the new CI processes run against the PR.

[IlyaKovalev]

Hello.
1 Done
2 Yes, i agree, we can implement this logic with PartitionRecord only if RecordPath will have hash functions and operations for number processing like mod. So PartitionRecord + RouteOnAttribute will have the same effect as this processor. (Implementation of weights can be awkward but it is definitely possible)
3 Renamed processor with DistributeHashRecord. I think the key feature here is distribution over hashed key.
and windows build failed ... hmm...

[alopresto]

I'm curious what this hash function is, and why there is a configurable property for it if there is only one allowable value.

[IlyaKovalev]

This is name of hash function for distribution. I suggest there will be several hash functions in future (MD5 for example). After keys hashing processor will decide where record should route (by division without remainder)

[alopresto]

Is it possible to add the other hashing algorithms now?

[IlyaKovalev]

Yes! Done!