Re: Apache Flink Operator State as Query Cache

2015-11-15 Thread Welly Tambunan
Hi Kostas, Yes. Exactly. Thanks a lot for this one. That's really what we need ! Cheers On Sun, Nov 15, 2015 at 8:53 PM, Kostas Tzoumas wrote: > Hi Wally, > > This version adds support for specifying and switching between time > semantics - processing time, ingestion time, or event time. > >

Different CoGroup behavior inside DeltaIteration

2015-11-15 Thread Truong Duc Kien
Hi, When running CoGroup between the solution set and a different dataset inside a DeltaIteration, the CoGroupFunction only get called for items that exist in the other dataset, simillar to an inner join. This is not the documented behavior for CoGroup: If a DataSet has a group with no matching

Custom TimestampExtractor and FlinkKafkaConsumer082

2015-11-15 Thread Konstantin Knauf
Hi everyone, I have the following issue with Flink (0.10) and Kafka. I am using a very simple TimestampExtractor like [1], which just extracts a millis timestamp from a POJO. In my streaming job, I read in these POJOs from Kafka using the FlinkKafkaConsumer082 like this: stream = env.addSource(n

Re: Apache Flink Operator State as Query Cache

2015-11-15 Thread Kostas Tzoumas
Hi Wally, This version adds support for specifying and switching between time semantics - processing time, ingestion time, or event time. When working with event time, you can specify watermarks to track the progress of event time. So, even if events arrive out of order, windows will be specified