[
https://issues.apache.org/jira/browse/KAFKA-4468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16095769#comment-16095769
]
Matthias J. Sax commented on KAFKA-4468:
----------------------------------------
[~guozhang] We have to distinguish two cases: (1) specifying a window in the
DSL via `TimeWindows` (plural). For this case, the window size is the same for
all windows within the store. (2) instantiating a new concrete window via `new
TimeWindow` (singular). This case could happen if someone uses PAPI with a
custom window store and custom window logic.
I am not sure how common the second case is, but if we want to support it we
would need to store the window size. Let me know if I am missing anything here.
Or do you think, that the second case could be covered by the session store?
> Correctly calculate the window end timestamp after read from state stores
> -------------------------------------------------------------------------
>
> Key: KAFKA-4468
> URL: https://issues.apache.org/jira/browse/KAFKA-4468
> Project: Kafka
> Issue Type: Bug
> Components: streams
> Reporter: Guozhang Wang
> Labels: architecture
>
> When storing the WindowedStore on the persistent KV store, we only use the
> start timestamp of the window as part of the combo-key as (start-timestamp,
> key). The reason that we do not add the end-timestamp as well is that we can
> always calculate it from the start timestamp + window_length, and hence we
> can save 8 bytes per key on the persistent KV store.
> However, after read it (via {{WindowedDeserializer}}) we do not set its end
> timestamp correctly but just read it as an {{UnlimitedWindow}}. We should fix
> this by calculating its end timestamp as mentioned above.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
