Question

假设我有这样的记录：

("a-b", "data1", 1)
("a-c", "data2", 1)
("a-b", "data3", 1)

如何在Apache Flink中进行分组和求和，以便得到以下结果？

("a-b", ["data1", "data3"], 2)
("a-c", ["data2"], 1)

此致凯文

Answer 1

我已在Flink shell（$FLINK_HOME/bin/start-scala-shell.sh local）中使用以下代码实现了此目的：

import org.apache.flink.util.Collector
benv.
  fromElements(("a-b", "data1", 1), ("a-c", "data2", 1), ("a-b", "data3", 1)).
  groupBy(0).
  reduceGroup { 
    (it: Iterator[(String, String, Int)], out: Collector[(String, List[String], Int)]) => {
      // Watch out: if the group is _very_ large this can lead to OOM errors
      val group = it.toList
      // For all groups with at least one element (prevent out-of-bounds)
      if (group.length > 0)
        // Get the "name", all the elements and their third-column aggregate
        out.collect((group(0)._1, group.map(_._2), group.map(_._3).sum))
    }
  }.print

使用以下输出

(a-b,List(data1, data3),2)
(a-c,List(data2),1)

Apache Flink - 求和并保持分组

1 个答案: