Hey, Yes that’s true for temporal table joins with event time. You would have to make sure that watermarks are progressing, like periodically publish them [1], or some other custom logic. I’m not sure what’s works out of the box in the Table API, however you can always register DataStream as a Table.
For the record, processing time temporal table joins emit updates immediately. [2] Piotrek [1] https://ci.apache.org/projects/flink/flink-docs-stable/dev/event_timestamps_watermarks.html#timestamp-assigners--watermark-generators <https://ci.apache.org/projects/flink/flink-docs-stable/dev/event_timestamps_watermarks.html#timestamp-assigners--watermark-generators> [2] https://ci.apache.org/projects/flink/flink-docs-stable/dev/table/streaming/joins.html#processing-time-temporal-joins <https://ci.apache.org/projects/flink/flink-docs-stable/dev/table/streaming/joins.html#processing-time-temporal-joins> > On 29 Jan 2020, at 11:48, Dominik Wosiński <wos...@gmail.com> wrote: > > Hey, > But isn't the temporal table join generating output only on watermark ??? I > have found such info here: > https://stackoverflow.com/questions/54441594/how-to-use-flink-temporal-tables. > But since one of the tables will have data that changes very rarely, this > would mean that using a temporal table would always wait for the slow > stream to create the watermark ?? > > Best Regards, > Dom.
