spark-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jungtaek Lim (JIRA)" <>
Subject [jira] [Commented] (SPARK-10816) EventTime based sessionization
Date Tue, 30 Oct 2018 05:34:00 GMT


Jungtaek Lim commented on SPARK-10816:

I feel we would be OK even we load sessions for a key into memory. Suppose the case which
accept late events by 30 days and session gap being set to 10 mins: at most 30 * 24 * 6 =
4,320 sessions can be made in given key. In practical case the number of sessions in a key
would be much smaller.

Anyway, I guess releasing 2.4.0 is the thing to concentrate on, so I would wait on release
and remind us again after we announce Spark 2.4.0. (Actually I have another patch which is
also waiting on releasing Spark 2.4.0.) I guess I explored most of options (except implementing
based on flatMapGroupsWithState) so we can revisit this when we have time, and check out the
result, and discuss how to move forward.

> EventTime based sessionization
> ------------------------------
>                 Key: SPARK-10816
>                 URL:
>             Project: Spark
>          Issue Type: New Feature
>          Components: Structured Streaming
>            Reporter: Reynold Xin
>            Priority: Major
>         Attachments: SPARK-10816 Support session window natively.pdf, Session Window
Support For Structure Streaming.pdf

This message was sent by Atlassian JIRA

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message