Re: Question about the process order in stream aggregate

Hi everybody, thank you all for your help. @Fabian I also check the DataStream that translated from the query and try to figure out what happens in each step. The results are as follows (correct me please if there's something wrong): Source -> Map (Order to Row3) -> FlatMap (do project and extra

Re: Possible bug in Kafka producer partitioning logic

That workaround should work, yes. The proper fix would also be something similar I guess, only just exposing extra APIs to properly provide different partitioners for different topics. Btw, sorry for the slow responses, as I’m also currently traveling for the Flink Forward conference in San Fran

[jira] [Created] (FLINK-6298) Local execution is not setting RuntimeContext for RichOutputFormat

Mateusz Zakarczemny created FLINK-6298: -- Summary: Local execution is not setting RuntimeContext for RichOutputFormat Key: FLINK-6298 URL: https://issues.apache.org/jira/browse/FLINK-6298 Project:

[jira] [Created] (FLINK-6297) CEP timeout does not trigger under certain conditions

Vijayakumar Palaniappan created FLINK-6297: -- Summary: CEP timeout does not trigger under certain conditions Key: FLINK-6297 URL: https://issues.apache.org/jira/browse/FLINK-6297 Project: Flink

[jira] [Created] (FLINK-6296) Retrieving execution plan fails succeeding execution attempt

Petr Novotnik created FLINK-6296: Summary: Retrieving execution plan fails succeeding execution attempt Key: FLINK-6296 URL: https://issues.apache.org/jira/browse/FLINK-6296 Project: Flink I

RE: Question about the process order in stream aggregate

Hi Xingcan, If you need to guarantee the order also in the case of procTime a trick that you can do is to set the working time of the env to processing time and to assign the proctime to the incoming stream. You can do this via .assignTimestampsAndWatermarks(new ...) And override override def

AW: Question about the process order in stream aggregate

Resending to dev@f.a.o Hi Xingcan, This is expected behavior. In general, is not possible to guarantee results for processing time. Your query is translated as follows: CollectionSrc(1) -round-robin-> MapFunc(n) -hash-part-> ProcessFunc(n) -fwd-> MapFunc(n) -fwd-> Sink(n) The order of record

RE: Question about the process order in stream aggregate

Hi Xingcan, Are you using parallelism 1 for the test? procTime semantics deals with the objects as they loaded in the operators. It could be the co-occuring partitioned events (in the same MS time frame) are processed in parallel and then the output is produced in different order. I suggest y

[jira] [Created] (FLINK-6295) use LoadingCache instead of WeakHashMap to lower latency

Tao Wang created FLINK-6295: --- Summary: use LoadingCache instead of WeakHashMap to lower latency Key: FLINK-6295 URL: https://issues.apache.org/jira/browse/FLINK-6295 Project: Flink Issue Type: Bug

[jira] [Created] (FLINK-6294) BucketingSink throws NPE while cancelling job

Andrey created FLINK-6294: - Summary: BucketingSink throws NPE while cancelling job Key: FLINK-6294 URL: https://issues.apache.org/jira/browse/FLINK-6294 Project: Flink Issue Type: Bug Compo

[jira] [Created] (FLINK-6293) Flakey JobManagerITCase

Nico Kruber created FLINK-6293: -- Summary: Flakey JobManagerITCase Key: FLINK-6293 URL: https://issues.apache.org/jira/browse/FLINK-6293 Project: Flink Issue Type: Bug Components: Job-S

[jira] [Created] (FLINK-6292) Travis: transfer.sh not accepting uploads via http:// anymore

Nico Kruber created FLINK-6292: -- Summary: Travis: transfer.sh not accepting uploads via http:// anymore Key: FLINK-6292 URL: https://issues.apache.org/jira/browse/FLINK-6292 Project: Flink Issu

[jira] [Created] (FLINK-6291) Internal Timer service cannot be "removed"

Gyula Fora created FLINK-6291: - Summary: Internal Timer service cannot be "removed" Key: FLINK-6291 URL: https://issues.apache.org/jira/browse/FLINK-6291 Project: Flink Issue Type: Bug