Spark operator: RDD key conversion operation (4) -cogroup, join

Cogroup

## The parameter is 1 RDD
Def cogroup [W] (other: RDD [(K, W)]): RDD [(K, (Iterable [V], Iterable [W]))
Def cogroup [W] (other: RDD [(K, W)], numPartitions: Int): RDD [(K, (Iterable [V], Iterable [W]))
Def cogroup [W] (other: RDD [(K, W)], partitioner: Partitioner): RDD [(K, (Iterable [V], Iterable [W]))
## parameters for 2 RDD
(K, (W, W1)], other2: RDD [(K, W2)]): RDD [(K, (Iterable [V], Iterable [W1], Iterable [W2] ])))
(K, W2)], numPartitions: Int): RDD [(K, (Iterable [V], Iterable [W1]], other words: , Iterable [W2]))]
(K, W2)], partitioner: Partitioner): RDD [(K, (Iterable [V], Iterable [W1]), other than Rc [W1, W2] , Iterable [W2]))]
## The parameter is 3 RDD
(K, W1)], other: RDD [(K, W2)], other3: RDD [(K, W3)]): RDD [(K, Iterable [V1], Iterable [W1], Iterable [W2], Iterable [W3]))
(C, W1)], other: RDD [(K, W2)], other3: RDD [(K, W3)], numPartitions: Int): RDD [ (Iterable [V], Iterable [W1], Iterable [W2], Iterable [W3]))
(C, W2)], other: RDD [(K, W1)], other2: RDD [(K, W2)], other3: RDD [(K, W3)], partitioner: Partitioner): RDD [ (Iterable [V], Iterable [W1], Iterable [W2], Iterable [W3]))
Cogroup is equivalent to SQL in the full outer association full outer join, return to the RDD in the record, the association is not on the empty.
Parameters numPartitions The number of partitions used to specify the result.
The parameter partitioner is used to specify the partition function.
## The argument is an example of RDD 
    

  1.   ("A", "1"), ("B", "2"), ("C", "3")), 2)
    2. 

  2.  ("A", "a"), ("C", "c"), ("D", "d")), 2)
    3. 

  3.  
    4. 

  4.  Scala> var rdd3 = rdd1.cogroup (rdd2)
    5. 

  5.  Rdd3: org.apache.spark.rdd.RDD [(String, (Iterable [String], Iterable [String])) = MapPartitionsRDD [12] at cogroup at: 25
    6. 

  6.  
    7. 

  7.  Scala> rdd3.partitions.size
    8. 

  8.  Res3: Int = 2
    9. 

  9.  
    10. 

  10.  Scala> rdd3.collect
    11. 

  11.  Res1: Array [(String, (Iterable [String], Iterable [String]))] = Array (
    12. 

  12.  (B, (CompactBuffer (2), CompactBuffer ())), 
    13. 

  13.  (D, (CompactBuffer (), CompactBuffer (d))), 
    14. 

  14.  (A, (CompactBuffer (1), CompactBuffer (a))), 
    15. 

  15.  (C, (CompactBuffer (3), CompactBuffer (c)))
    16. 

  16.  )
    17. 

  17.  
    18. 

  18.  
    19. 

  19.  Scala> var rdd4 = rdd1.cogroup (rdd2,3)
    20. 

  20.  Rdd4: org.apache.spark.rdd.RDD [(String, (Iterable [String], Iterable [String])) = MapPartitionsRDD [14] at cogroup at: 25
    21. 

  21.  
    22. 

  22.  Scala> rdd4.partitions.size
    23. 

  23.  Res5: Int = 3
    24. 

  24.  
    25. 

  25.  Scala> rdd4.collect
    26. 

  26.  Res6: Array [(String, (Iterable [String], Iterable [String]))] = Array (
    27. 

  27.  (B, ( CompactBuffer (2) , CompactBuffer ())), 
    28. 

  28.  (C, ( CompactBuffer (3) , CompactBuffer (c))), 
    29. 

  29.  (A, ( CompactBuffer (1) , CompactBuffer (a))), 
    30. 

  30.  (D, ( CompactBuffer () , CompactBuffer (d))))
    31. 

  31.  
    32.
## The argument is an example of two RDDs 
    

  1.   ("A", "1"), ("B", "2"), ("C", "3")), 2)
    2. 

  2.  ("A", "a"), ("C", "c"), ("D", "d")), 2)
    3. 

  3.  Var rdd3 = sc.makeRDD (Array ("A", "A"), ("E", "E")), 2)
    4. 

  4.  
    5. 

  5.  Scala> var rdd4 = rdd1.cogroup (rdd2, rdd3)
    6. 

  6.  Rdd4: org.apache.spark.rdd.RDD [(String, (Iterable [String], Iterable [String], Iterable [String]))] = 
    7. 

  7.  MapPartitionsRDD [17] at cogroup at: 27
    8. 

  8.  
    9. 

  9.  Scala> rdd4.partitions.size
    10. 

  10.  Res7: Int = 2
    11. 

  11.  
    12. 

  12.  Scala> rdd4.collect
    13. 

  13.  Res9: Array [(String, (Iterable [String], Iterable [String], Iterable [String]))] = Array (
    14. 

  14.  (B, ( CompactBuffer (2) , CompactBuffer (), CompactBuffer () )), 
    15. 

  15.  (D, ( CompactBuffer () , CompactBuffer (d), CompactBuffer () )), 
    16. 

  16.  (A, ( CompactBuffer (1) , CompactBuffer (a), CompactBuffer (A) )), 
    17. 

  17.  ( CompactBuffer (3) , CompactBuffer (c), CompactBuffer ( )), 
    18. 

  18.  (E, ( CompactBuffer () , CompactBuffer (), CompactBuffer (E) )))
    19. 

  19.  
    20. 

  20.  
    21.
## parameters for the three RDD examples slightly, ibid.

Join

Def join [W] (other: RDD [(K, W)]): RDD [(K, (V, W))]
Def join [W] (other: RDD [(K, W)], numPartitions: Int): RDD [(K, (V, W))]
Def join [W] (other: RDD [(K, W)], partitioner: Partitioner): RDD [(K, (V, W))]
Join is equivalent to the SQL in the context of the join, only two RDD According to the K can be associated with the results, join can only be used for the association between the two RDD, if more than one RDD association, many times can be related.
Parameters numPartitions The number of partitions used to specify the result
The parameter partitioner is used to specify the partition function 
    

  1.   ("A", "1"), ("B", "2"), ("C", "3")), 2)
    2. 

  2.  ("A", "a"), ("C", "c"), ("D", "d")), 2)
    3. 

  3.  
    4. 

  4.  Scala> rdd1.join (rdd2) .collect
    5. 

  5.  Res10: Array [(String, (String, String))] = Array ((A, (1, a)), (C, (3, c)))
    6. 

  6.  
    7.
For more information on the Spark operator, refer to the Spark operator series .
Http://lxw1234.com/archives/2015/07/363.htm

Commentaires

Enregistrer un commentaire

Posts les plus consultés de ce blog

Controlling Parallelism in Spark by controlling the input partitions by controlling the input partitions

Image
It is useful to be able to control the degree of parallelism when using Spark. Spark provides a very convenient method to increase the degree of parallelism which should be adequate in practice. This blog entry of the "Inside Spark" series describes the knobs Spark uses to control the degree of parallelism. Controlling the number of partitions in the " local " mode Spark is designed ground up to enable unit testing. To control the degree of parallelism in the local mode simply utilize the function JavaRDD<String> org.apache.spark.api.java.JavaSparkContext.parallelize(List<String> list, int numSlices) The following snippet of code demonstrates this usage- public class SparkControlPartitionSizeLocally { public static void main ( String [] args ) { JavaSparkContext sc = new JavaSparkContext ( "local" , "localpartitionsizecontrol" ); String input = "four score and sev
Lire la suite

Spark performance optimization: shuffle tuning

Image
Article directory 1 shuffle tuning 1.1 Summary of tuning 1.2 Overview of ShuffleManager Development 1.3 HashShuffleManager operating principle 1.3.1 Unoptured HashShuffleManager 1.3.2 Optimized HashShuffleManager 1.4 SortShuffleManager operating principle 1.4.1 General operating mechanism 1.4.2 bypass running mechanism 1.5 shuffle related parameters tuning 1.5.1 spark.shuffle.file.buffer 1.5.2 spark.reducer.maxSizeInFlight 1.5.3 spark.shuffle.io.maxRetries 1.5.4 spark.shuffle.io.retryWait 1.5.5 spark.shuffle.memoryFraction 1.5.6 spark.shuffle.manager 1.5.7 spark.shuffle.sort.bypassMergeThreshold 1.5.8 spark.shuffle.consolidateFiles 2 write in the last words Shuffle Summary of tuning Most of the performance of Spark operations is mainly consumed in the shuffle link, because the link contains a large number of disk IO, serialization, network data transmission and othe
Lire la suite

Spark optimization

Image
One operation and maintenance 1. Master hang up, standby restart is also invalid Master defaults to 512M of memory, when the task in the cluster is particularly high, it will hang, because the master will read each task event log log to generate spark ui, the memory will naturally OOM, you can run the log See that the master of the start through the HA will naturally fail for this reason. solve Increase the Master's memory spark-env.sh , set in the master node spark-env.sh : export SPARK_DAEMON_MEMORY 10g # 根据你的实际情况 Reduce the job information stored in the Master memory spark.ui.retainedJobs 500 # 默认都是1000 spark.ui.retainedStages 500 Hang up or suspend Sometimes we will see the web node in the web ui disappear or in the dead state, the task of running the node will report a variety of lost worker errors, causing the same reasons and the above, worker memory to save a lot of ui The information leads to gc when the heartbeat is lost
Lire la suite