Github user rtudoran commented on the issue: https://github.com/apache/flink/pull/4380 @fhueske I did not understood initially that this is your suggestion. What you propose has the advantage that it is easy to maintain (as we consolidate the whole functionality) and a slight disadvantage that you have a couple of useless checks in some scenarios (e.g., you would still check the fetch condition even if you would have only offset or just the plain sorting). If this tiny performance price to pay is ok ...than clearly we can consolidate the implementation. Nevertheless, even if this is the case i would still propose we keep the ProcTimeIdentitySortProcessFunction This is for the scenario where you have the order by simply on the time (and no other field). In the case of simple sort we only had an identity map function to pass the events. For the case of offset/fetch we can either extend that or keep one different implementation that adds the state counters for offset/fetch. I propose we keep this later implementation.
--- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---