返回类型为复杂对象数组的Scala UDAF

问题描述

我是Scala开发的新手，正在尝试解决以下问题:

I am new to Scala development and trying to solve below issue:

我有一个UDAF，它返回复杂对象的数组(它是一个字符串和字符串数组).在更新方法中，缓冲区返回的是wrappedArray类型，我不知道如何使用缓冲区中的新值进行更新.我曾尝试将其转换为seq，但无法正常工作...

I have a UDAF which returns array of complex object ( which is a string and array of strings). In the update method buffer is returning of type wrappedArray and I don't know how to update back with new value in buffer. I have tried converting it to seq but didn't work...

case class variablePairs(val variable1: String, val Respondents: Seq[String])

    import java.util
    import java.util.Collections

    import org.apache.spark.sql.Row
    import org.apache.spark.sql.expressions.MutableAggregationBuffer
    import org.apache.spark.sql.expressions.UserDefinedAggregateFunction
    import org.apache.spark.sql.types.{DataType, DataTypes, StringType, StructType}

    class MyUDF extends UserDefinedAggregateFunction {

      override def inputSchema(): StructType =
        new StructType()
          .add("variable1", DataTypes.StringType)
          .add("variable2CSList", DataTypes.StringType)

      //intermediate schema
      override def bufferSchema(): StructType =
        new StructType()
          .add("Households", DataTypes.createArrayType(
          DataTypes.StringType))
            new StructType()
              .add("variable1",DataTypes.StringType)
              .add("variable2",DataTypes.createArrayType(
                DataTypes.StringType
              ))


      //output schema
      override def dataType(): DataType =   new StructType()
          new StructType()
            .add("Households", DataTypes.createArrayType(
              new StructType()
                .add("variable1",DataTypes.StringType)
                .add("variable2",DataTypes.createArrayType(
                  DataTypes.StringType
                ))))


      override def deterministic(): Boolean = true

      override def initialize(buffer: MutableAggregationBuffer): Unit = {
        buffer.update(0,  Seq[String]()) 
      }

      override def update(buffer: MutableAggregationBuffer, row: Row): Unit = {

        val variable1: String = row.getString(0)
        val variable2CSList:String = row.getString(1);

        val respondentsIdArray:Array[String] = variable2CSList.split(",")

        val houseHold:variablePairs = variablePairs(variable1 = variable1, Respondents = respondentsIdArray.toSeq )

        val wrappedArray =  buffer.get(0).asInstanceOf[Seq[String]]

        val households:Seq[variablePairs] = Seq(houseHold)

        buffer.update(0,wrappedArray.toArray ++ variable1)
      }

      override def merge(buffer: MutableAggregationBuffer, row: Row): Unit = {
        val oldList = buffer.getList[variablePairs](0);
        val newList = row.getList[variablePairs](0);

         buffer.update(0,oldList.addAll(newList))
      }

      override def evaluate(row: Row): AnyRef = {
        new Tuple1(row.get(0));
      }
    }

    I got below error while running this code:

App > 17/10/10 22:01:48 task-result-getter-1 WARN TaskSetManager: Lost task 2.0 in stage 1.0 (TID 31, ip-10-61-41-163.ec2.internal, executor 3): scala.MatchError: 1 (of class java.lang.Character)
App > at org.apache.spark.sql.catalyst.CatalystTypeConverters$StringConverter$.toCatalystImpl(CatalystTypeConverters.scala:276)
App > at org.apache.spark.sql.catalyst.CatalystTypeConverters$StringConverter$.toCatalystImpl(CatalystTypeConverters.scala:275)
App > at org.apache.spark.sql.catalyst.CatalystTypeConverters$CatalystTypeConverter.toCatalyst(CatalystTypeConverters.scala:103)
App > at org.apache.spark.sql.catalyst.CatalystTypeConverters$ArrayConverter$$anonfun$toCatalystImpl$1.apply(CatalystTypeConverters.scala:162)
App > at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:234)
App > at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:234)
App > at scala.collection.IndexedSeqOptimized$class.foreach(IndexedSeqOptimized.scala:33)
App > at scala.collection.mutable.ArrayOps$ofRef.foreach(ArrayOps.scala:186)
App > at scala.collection.TraversableLike$class.map(TraversableLike.scala:234)
App > at scala.collection.mutable.ArrayOps$ofRef.map(ArrayOps.scala:186)
App > at org.apache.spark.sql.catalyst.CatalystTypeConverters$ArrayConverter.toCatalystImpl(CatalystTypeConverters.scala:162)
App > at org.apache.spark.sql.catalyst.CatalystTypeConverters$ArrayConverter.toCatalystImpl(CatalystTypeConverters.scala:154)
App > at org.apache.spark.sql.catalyst.CatalystTypeConverters$CatalystTypeConverter.toCatalyst(CatalystTypeConverters.scala:103)
App > at org.apache.spark.sql.catalyst.CatalystTypeConverters$$anonfun$createToCatalystConverter$2.apply(CatalystTypeConverters.scala:383)
App > at org.apache.spark.sql.execution.aggregate.MutableAggregationBufferImpl.update(udaf.scala:246)
App > at com.turner.audiencematters.udf.RespondentPairUDF.update(RespondentPairUDF.scala:65)
App > at org.apache.spark.sql.execution.aggregate.ScalaUDAF.update(udaf.scala:425)
App > at org.apache.spark.sql.execution.aggregate.AggregationIterator$$anonfun$1$$anonfun$applyOrElse$1.apply(AggregationIterator.scala:171)
App > at org.apache.spark.sql.execution.aggregate.AggregationIterator$$anonfun$1$$anonfun$applyOrElse$1.apply(AggregationIterator.scala:171)
App > at org.apache.spark.sql.execution.aggregate.AggregationIterator$$anonfun$generateProcessRow$1.apply(AggregationIterator.scala:187)
App > at org.apache.spark.sql.execution.aggregate.AggregationIterator$$anonfun$generateProcessRow$1.apply(AggregationIterator.scala:181)
App > at org.apache.spark.sql.execution.aggregate.SortBasedAggregationIterator.processCurrentSortedGroup(SortBasedAggregationIterator.scala:122)
App > at org.apache.spark.sql.execution.aggregate.SortBasedAggregationIterator.next(SortBasedAggregationIterator.scala:157)
App > at org.apache.spark.sql.execution.aggregate.SortBasedAggregationIterator.next(SortBasedAggregationIterator.scala:29)
App > at scala.collection.Iterator$$anon$11.next(Iterator.scala:409)
App > at org.apache.spark.shuffle.sort.BypassMergeSortShuffleWriter.write(BypassMergeSortShuffleWriter.java:150)
App > at org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:96)
App > at org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:53)
App > at org.apache.spark.scheduler.Task.run(Task.scala:99)
App > at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:322)
App > at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
App > at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
App > at java.lang.Thread.run(Thread.java:745)

推荐答案

我有一个类似的问题，尽管只是尝试返回一个 Array [String]

I had a similar problem though only trying to return an Array[String]

这段代码很有帮助:

https://gist.github.com/sadikovi/7608c8c7eb5d7fe69a1a

摘录了在我的UDAF中对我有用的代码:

Extracts of the code which worked for me in my UDAF:

...
override def dataType: DataType = ArrayType(StringType)
...
override def evaluate(buffer: Row): Array[String] = {
...

我希望这会有所帮助！

这篇关于返回类型为复杂对象数组的Scala UDAF的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持！