使用 sc.union 时出现 Spark 堆栈溢出错误

Question

我正在使用 spark 的 scala api 并尝试合并 3000 个 RDD。

我正在做一个 for 循环，迭代 3000 个文本文件，创建一个 RDD 并在旧 RDD 和刚刚从文本文件创建的新 RDD 中建立一个联合。

这是我的代码：

var tempRDD = sc.textFile(s"" + fileNameArray(ind)).filter(_.split(",",-1).length == cols).filter(line => line.split(",").filter(cell => cell.length > 0).length == cols).
    map{line => 
        val parts = line.split(",",-1);
        val relaventFeatures = Array(parts(ind0),parts(ind1),parts(ind2));
        val vector = Vectors.dense(relaventFeatures.map(_.toDouble));
        val fileName = fileNameArray(ind);
        (fileName, (parts(0), vector))};
    if ( !tempRDD.isEmpty){
        allRDD = sc.union(allRDD, tempRDD);
    }

我看到question 询问有关在使用 RDD.union 时出现堆栈溢出的问题，答案说如果使用 SparkContext.union 不会发生堆栈溢出 - 但我得到了一个：

java.lang.StackOverflowError at org.apache.spark.rdd.UnionRDD$$anonfun$1.apply(UnionRDD.scala:66) at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) at scala.collection.immutable.List.foreach(List.scala:318) at scala.collection.TraversableLike$class.map(TraversableLike.scala:244) at scala.collection.AbstractTraversable.map(Traversable.scala:105) at org.apache.spark.rdd.UnionRDD.getPartitions(UnionRDD.scala:66) at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:239) at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:237) at scala.Option.getOrElse(Option.scala:120) at org.apache.spark.rdd.RDD.partitions(RDD.scala:237) at org.apache.spark.rdd.UnionRDD$$anonfun$1.apply(UnionRDD.scala:66) at org.apache.spark.rdd.UnionRDD$$anonfun$1.apply(UnionRDD.scala:66) at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) at scala.collection.immutable.List.foreach(List.scala:318) at scala.collection.TraversableLike$class.map(TraversableLike.scala:244) at scala.collection.AbstractTraversable.map(Traversable.scala:105) at org.apache.spark.rdd.UnionRDD.getPartitions(UnionRDD.scala:66) at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:239) at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:237) at scala.Option.getOrElse(Option.scala:120) at org.apache.spark.rdd.RDD.partitions(RDD.scala:237) at org.apache.spark.rdd.UnionRDD$$anonfun$1.apply(UnionRDD.scala:66) at org.apache.spark.rdd.UnionRDD$$anonfun$1.apply(UnionRDD.scala:66) at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) at scala.collection.immutable.List.foreach(List.scala:318) at scala.collection.TraversableLike$class.map(TraversableLike.scala:244) at scala.collection.AbstractTraversable.map(Traversable.scala:105) at org.apache.spark.rdd.UnionRDD.getPartitions(UnionRDD.scala:66) at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:239) at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:237) at scala.Option.getOrElse(Option.scala:120) at org.apache.spark.rdd.RDD.partitions(RDD.scala:237) at org.apache.spark.rdd.UnionRDD$$anonfun$1.apply(UnionRDD.scala:66) at org.apache.spark.rdd.UnionRDD$$anonfun$1.apply(UnionRDD.scala:66) at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) at scala.collection.immutable.List.foreach(List.scala:318) at scala.collection.TraversableLike$class.map(TraversableLike.scala:244) at scala.collection.AbstractTraversable.map(Traversable.scala:105) at org.apache.spark.rdd.UnionRDD.getPartitions(UnionRDD.scala:66) at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:239) at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:237) at scala.Option.getOrElse(Option.scala:120) at org.apache.spark.rdd.RDD.partitions(RDD.scala:237) at org.apache.spark.rdd.UnionRDD$$anonfun$1.apply(UnionRDD.scala:66) at org.apache.spark.rdd.UnionRDD$$anonfun$1.apply(UnionRDD.scala:66) at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) at scala.collection.immutable.List.foreach(List.scala:318) at scala.collection.TraversableLike$class.map(TraversableLike.scala:244) at scala.collection.AbstractTraversable.map(Traversable.scala:105) at org.apache.spark.rdd.UnionRDD.getPartitions(UnionRDD.scala:66) at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:239) at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:237) at scala.Option.getOrElse(Option.scala:120) at org.apache.spark.rdd.RDD.partitions(RDD.scala:237) at org.apache.spark.rdd.UnionRDD$$anonfun$1.apply(UnionRDD.scala:66) at org.apache.spark.rdd.UnionRDD$$anonfun$1.apply(UnionRDD.scala:66) at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) at scala.collection.immutable.List.foreach(List.scala:318) at scala.collection.TraversableLike$class.map(TraversableLike.scala:244) at scala.collection.AbstractTraversable.map(Traversable.scala:105) at org.apache.spark.rdd.UnionRDD.getPartitions(UnionRDD.scala:66) at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:239) at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:237) at scala.Option.getOrElse(Option.scala:120) at org.apache.spark.rdd.RDD.partitions(RDD.scala:237) at org.apache.spark.rdd.UnionRDD$$anonfun$1.apply(UnionRDD.scala:66) at org.apache.spark.rdd.UnionRDD$$anonfun$1.apply(UnionRDD.scala:66) at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) at scala.collection.immutable.List.foreach(List.scala:318) at scala.collection.TraversableLike$class.map(TraversableLike.scala:244) at scala.collection.AbstractTraversable.map(Traversable.scala:105) at org.apache.spark.rdd.UnionRDD.getPartitions(UnionRDD.scala:66) at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:239) at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:237) at scala.Option.getOrElse(Option.scala:120) at org.apache.spark.rdd.RDD.partitions(RDD.scala:237) at org.apache.spark.rdd.UnionRDD$$anonfun$1.apply(UnionRDD.scala:66) at org.apache.spark.rdd.UnionRDD$$anonfun$1.apply(UnionRDD.scala:66) at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) at scala.collection.immutable.List.foreach(List.scala:318) at scala.collection.TraversableLike$class.map(TraversableLike.scala:244) at scala.collection.AbstractTraversable.map(Traversable.scala:105) at org.apache.spark.rdd.UnionRDD.getPartitions(UnionRDD.scala:66) at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:239) at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:237) at scala.Option.getOrElse(Option.scala:120) at org.apache.spark.rdd.RDD.partitions(RDD.scala:237) at org.apache.spark.rdd.UnionRDD$$anonfun$1.apply(UnionRDD.scala:66) at org.apache.spark.rdd.UnionRDD$$anonfun$1.apply(UnionRDD.scala:66) at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) at scala.collection.immutable.List.foreach(List.scala:318) at scala.collection.TraversableLike$class.map(TraversableLike.scala:244) at scala.collection.AbstractTraversable.map(Traversable.scala:105) at org.apache.spark.rdd.UnionRDD.getPartitions(UnionRDD.scala:66) at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:239) at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:237) at scala.Option.getOrElse(Option.scala:120) at org.apache.spark.rdd.RDD.partitions(RDD.scala:237) at org.apache.spark.rdd.UnionRDD$$anonfun$1.apply(UnionRDD.scala:66) at org.apache.spark.rdd.UnionRDD$$anonfun$1.apply(UnionRDD.scala:66) at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) at scala.collection.immutable.List.foreach(List.scala:318) at scala.collection.TraversableLike$class.map(TraversableLike.scala:244) at scala.collection.AbstractTraversable.map(Traversable.scala:105) at org.apache.spark.rdd.UnionRDD.getPartitions(UnionRDD.scala:66) at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:239) at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:237) at scala.Option.getOrElse(Option.scala:120) at org.apache.spark.rdd.RDD.partitions(RDD.scala:237) at org.apache.spark.rdd.UnionRDD$$anonfun$1.apply(UnionRDD.scala:66) at org.apache.spark.rdd.UnionRDD$$anonfun$1.apply(UnionRDD.scala:66) at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) at scala.collection.immutable.List.foreach(List.scala:318) at scala.collection.TraversableLike$class.map(TraversableLike.scala:244) at scala.collection.AbstractTraversable.map(Traversable.scala:105) at org.apache.spark.rdd.UnionRDD.getPartitions(UnionRDD.scala:66) at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:239) at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:237) at scala.Option.getOrElse(Option.scala:120) at org.apache.spark.rdd.RDD.partitions(RDD.scala:237) at org.apache.spark.rdd.UnionRDD$$anonfun$1.apply(UnionRDD.scala:66) at org.apache.spark.rdd.UnionRDD$$anonfun$1.apply(UnionRDD.scala:66) at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) at scala.collection.immutable.List.foreach(List.scala:318) at scala.collection.TraversableLike$class.map(TraversableLike.scala:244) at scala.collection.AbstractTraversable.map(Traversable.scala:105) at org.apache.spark.rdd.UnionRDD.getPartitions(UnionRDD.scala:66) at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:239) at org.apache.spark.rdd.RDD$$anonfun$partitions$2.apply(RDD.scala:237) at scala.Option.getOrElse(Option.scala:120)

有人知道我做错了什么吗？

Answer 1

首先。您的 allRdd 似乎是一个 var。尽量避免这种情况；你可以这样做：

val allRdd = fileNameArray.map(name => {
  val rdd = sc.textFile(s"" + name).filter(...)
}).reduce(_ + _)

现在...加入 3000 个 RDD 听起来有点疯狂。联合方法使这个：

new UnionRDD(sc, Array(this, other))

如果你加入 3000 个 RDD，你会得到：

new UnionRDD(sc, Array(this, new UnionRDD(sc, Array(this2, new UnionRDD(sc, Array(this3, new UnionRDD(sc, Array(this4, (........)))))))))

这就是你的 StackOverflow

的原因

您可以使用通配符并从目录中读取它们...

val allRdds = sc.textFile("filesDir/*.txt"). // or whatever
filter(...)

或者，如果您更喜欢阅读文件列表：

val allRdds = sc.textFile(fileNameArray.mkString(",")).filter(...)

更新：

由于你需要文件名，你可以使用wholeTextFiles函数：

val files: RDD[(String, String)] = sc.wholeTextFiles(fileNameArray.mkString(",")).
  flatMap(f => f._2.split("\n").map(l => (f._1, l) ) )
files.filter(...).mapValues(...)