全网最详细4W字Flink入门笔记（上） 5

val env = StreamExecutionEnvironment.getExecutionEnvironment
val stream = env.generateSequence(1,10).setParallelism(1)
println(stream.getParallelism)
stream.shuffle.print()
env.execute()

2> 1
1> 4
7> 10
4> 6
6> 3
5> 7
8> 2
1> 5
1> 8
1> 9

val env = StreamExecutionEnvironment.getExecutionEnvironment
env.setParallelism(3)
val stream = env.generateSequence(1,100)
val shuffleStream = stream.rebalance
shuffleStream.print()
env.execute()

8> 6
3> 1
5> 3
7> 5
1> 7
2> 8
6> 4
4> 2
3> 9
4> 10

val env = StreamExecutionEnvironment.getExecutionEnvironment
val stream = env.generateSequence(1,10).setParallelism(2)
stream.writeAsText("./data/stream1").setParallelism(2)
stream.rescale.writeAsText("./data/stream2").setParallelism(4)
env.execute()

1
3
5
7
9

2
4
6
8
10

1
5
9

3
7

2
6
10

4
8

val env = StreamExecutionEnvironment.getExecutionEnvironment
val stream = env.generateSequence(1,10).setParallelism(2)
stream.writeAsText("./data/stream1").setParallelism(2)
stream.broadcast.writeAsText("./data/stream2").setParallelism(4)
env.execute()

1
3
5
7
9

2
4
6
8
10

1
3
5
7
9
2
4
6
8
10

val env = StreamExecutionEnvironment.getExecutionEnvironment
val stream = env.generateSequence(1,10).setParallelism(2)
stream.writeAsText("./data/stream1").setParallelism(2)
stream.global.writeAsText("./data/stream2").setParallelism(4)
env.execute()

1
3
5
7
9

2
4
6
8
10

1
3
5
7
9
2
4
6
8
10

Forward partitioning does not allow change of parallelism
* Upstream operation: Source: Sequence Source-1 parallelism: 2,
* downstream operation: Sink: Unnamed-4 parallelism: 4
* stream.forward.writeAsText("./data/stream2").setParallelism(4)

val env = StreamExecutionEnvironment.getExecutionEnvironment
val stream = env.generateSequence(1,10).setParallelism(2)
stream.writeAsText("./data/stream1").setParallelism(2)
stream.forward.writeAsText("./data/stream2").setParallelism(2)
env.execute()

1
3
5
7
9

2
4
6
8
10

1
3
5
7
9

2
4
6
8
10

MathUtils.murmurHash(keyHash)（每个元素的Hash值） % maxParallelism（下游分区数）

val env = StreamExecutionEnvironment.getExecutionEnvironment
val stream = env.generateSequence(1,10).setParallelism(2)
stream.writeAsText("./data/stream1").setParallelism(2)
stream.keyBy(0).writeAsText("./data/stream2").setParallelism(2)
env.execute()

object ShuffleOperator {
  def main(args: Array[String]): Unit = {
    val env = StreamExecutionEnvironment.getExecutionEnvironment
    env.setParallelism(2)
    val stream = env.generateSequence(1,10).map((_,1))
    stream.writeAsText("./data/stream1")
    stream.partitionCustom(new customPartitioner(),0)
      .writeAsText("./data/stream2").setParallelism(4)
    env.execute()
  }
  class customPartitioner extends Partitioner[Long]{
    override def partition(key: Long, numPartitions: Int): Int = {
      key.toInt % numPartitions
    }
  }
}