Hi, I was trying to implement a better way to handle data skew using Flink and I found this talk from #FlinkForward SF 2017: "Cliff Resnick & Seth Wiesman - From Zero to Streaming <https://youtu.be/mSLesPzWplA?t=835>" [1] which says that they used OneInputStreamOperator [2]. Through it, they could implement the "combiner" in Hadoop (execute part of the reduce tasks on the Map phase, before shuffling).
I need some help here. What are some of the Flink source-code operators that I can peek up to implement my on operator that deals with data skew? Or maybe, is there someone that have an example of a use case similar to this? [1] https://youtu.be/mSLesPzWplA?t=835 [2] https://ci.apache.org/projects/flink/flink-docs-master/api/java/index.html?org/apache/flink/streaming/api/functions/source/ContinuousFileReaderOperator.html Thanks! Felipe *--* *-- Felipe Gutierrez* *-- skype: felipe.o.gutierrez* *--* *https://felipeogutierrez.blogspot.com <https://felipeogutierrez.blogspot.com>*
