如何设置函数文本的输出类型?

Row 方法 toSeq 返回一个 Seq[Any] ,使用 getAs[T](index) 如下图所示:

import org.apache.spark.sql.Row

val df = Seq(
  (1, "a", 1.0),
  (2, "b", 2.0)
).toDF("c1", "c2", "c3")

val rdd = df.rdd
// rdd: org.apache.spark.rdd.RDD[org.apache.spark.sql.Row] = ...

rdd.map(_.getAs[Int](0))
// res1: org.apache.spark.rdd.RDD[Int]

rdd.map(_.getAs[String](1))
// res2: org.apache.spark.rdd.RDD[String]

rdd.map(_.getAs[Double](2))
// res3: org.apache.spark.rdd.RDD[Double]

回答标题中的实际问题:使用类型归属。

scala>  val foo = (i: Int) => Set.empty
foo: Int => scala.collection.immutable.Set[Nothing] = $$Lambda$funnyNumber1

归属:

scala> val foo = (i: Int) => (Set.empty: Set[Int])
foo: Int => Set[Int] = $$Lambda$funnyNumber2