Hi Greg. Based on a quick test I cannot reproduce the issue, it is emitting messages correctly in the ITCase environment. can you share more information? Does the same problem happen if you use proctime? I am guessing this could be highly correlated with how you set your watermark strategy of your input streams of "user_things" and "user_stuff".
-- Rong On Tue, Jun 26, 2018 at 6:37 PM Gregory Fee <g...@lyft.com> wrote: > Hello User Community! > > I am running some streaming SQL that involves a union all into an over > window similar to the below: > > SELECT user_id, count(action) OVER (PARTITION BY user_id ORDER BY rowtime > RANGE INTERVAL '14' DAY PRECEDING) action_count, rowtime > FROM > (SELECT rowtime, user_id, thing as action FROM user_things > UNION ALL SELECT rowtime, user_id, stuff as action FROM user_stuff) > > The SQL generates three operators. There are two operators that process > the 'from' part of the clause that feed into an 'over' operator. I notice > that messages flow into the 'over' operator and just buffer there for a > long time (hours in some cases). Eventually something happens and the data > starts to flush through to the downstream operators. Can anyone help me > understand what is causing that behavior? I want the data to flow through > more consistently. > > Thanks! > > > > -- > *Gregory Fee* > Engineer > 425.830.4734 <+14258304734> > [image: Lyft] <http://www.lyft.com> >
