That's what I thought Dian.
The problem is that setting the watermark strategy like that didn't work
either, the method on_event_time is never called.
I did some reading of
https://ci.apache.org/projects/flink/flink-docs-release-1.13/docs/dev/datastream/event-time/generating_watermarks/#watermark
Hi Ignacio,
Yes, you are right that you need to define the watermark strategy explicitly in
case of event time processing.
Regarding to *with_timestamp_assigner*, this is optional. If you don’t define
it, it will generate watermark according to the timestamp extracted from the
Kafka record (Co
I assumed that the event time and watermarks were already handled by the
Kafka connector.
So, basically, I need to do something like:
stream.assign_timestamps_and_watermarks(
WatermarkStrategy.for_bounded_out_of_orderness(Duration.of_seconds(30))
)
Do I also need to set the timestamps myself by
I'm trying to so a simple word count example, here's the main function:
def main():
parser = argparse.ArgumentParser()
parser.add_argument('--kafka-clients-jar', required=True)
parser.add_argument('--flink-connector-kafka-jar', required=True)
args = parser.parse_args()
env =
Regarding "Kafka consumer doesn’t read any message”: I’m wondering about this.
Usually the processing logic should not affect the Kafka consumer. Did you
judge this as there is no output for the job? If so, I’m guessing that it’s
because the window wasn’t triggered in case of event-time.
Could
