Github user sunjincheng121 commented on the issue:
https://github.com/apache/flink/pull/3629
Hi @fhueske thanks a lot for your review. I have updated the PR according
to your comments.
There is one thing need your explanation.i.e. :
`Also, I realized that this implementation (other OVER windows are probably
affected as well) will not discard state if the key space evolves. We should
add a JIRA to add a configuration parameter to remove state if no row was
received for a certain amount of time.`
IMO. For event-time case, we use data-driven management state. each
processing a data will be timely processing of expired data. if the data
continuously, our state will be promptly processed. If there is no data for a
long time, the data in the state will not expand. If the next data is not sure
when the arrival of the case, I think we should not clear the data, because the
removal of data will lead to the next calculation error. If the state data has
TTL settings, user can config TTL which can be friendly to clear the state. If
i understand you correctly you said that the configuration parameters, that is
the the TTL config, is this correct? If not so, I'm appreciated If you can tell
me your detailed thoughts.
Thanks,
SunJincheng
---
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 [email protected] or file a JIRA ticket
with INFRA.
---
