Hi, I'm afraid that's not possible but you can use a regular stream and do the join yourself. What the code for JoinedStreams essentially does is take two streams, map them to a common data type, union them and then perform a normal window operation.
The code for this is in CoGroupedStreams (as the general case of a join) and JoinedStreams. Cheers, Aljoscha On Mon, 30 Jan 2017 at 17:38 Saiph Kappa <saiph.ka...@gmail.com> wrote: > Hi all, > > Is it possible to specify allowed lateness for a window join like the > following one: > > val tweetsAndWarning = > warningsPerStock.join(tweetsPerStock).where(_.symbol).equalTo(_.symbol) > .window(SlidingEventTimeWindows.of(Time.of(windowDurationSec, > TimeUnit.SECONDS), Time.of(windowDurationSec, > TimeUnit.SECONDS))) > .apply((c1, c2) => (c1.count, c2.count)) > > > I think it is related with these: > > https://cwiki.apache.org/confluence/display/FLINK/Streaming+Window+Join+Rework > https://issues.apache.org/jira/browse/FLINK-3109 > > > Thanks! >
