Spark 2.x Scala 2.1.1自定义编码器类类型不匹配 [英] Spark 2.x scala 2.1.1 custom encoder class type mismatch

问题描述

我正在使用此代码在scala 2.1.1中为spark 2.x定义编码器:

I am using this code to define an encoder for spark 2.x, in scala 2.1.1:

import org.apache.spark.sql.Encoder
import org.apache.spark.sql.types.StructType

class LogProcessorMessageEncoder extends Encoder[LogProcessorMessage] {
    override def schema: StructType = null

    override def clsTag: Class[LogProcessorMessage] = classOf[LogProcessorMessage]
}

object LogProcessorMessage {}

class LogProcessorMessage extends Serializable {}

并且IntelliJ说很好，但是对clsTag方法的重写会导致编译器错误:

And IntelliJ says it is fine, but the override of the clsTag method results in a compiler error:

Error:(20, 15) overriding method clsTag in trait Encoder of type => scala.reflect.ClassTag[dvn.aast.telemetry.scala.stream.LogProcessorMessage];
method clsTag has incompatible type
override def clsTag: Class[LogProcessorMessage] = classOf[LogProcessorMessage]

推荐答案

clsTag 必须是 ClassTag 而不是 Class :

class LogProcessorMessageEncoder extends Encoder[LogProcessorMessage] {
  override def clsTag: ClassTag[LogProcessorMessage] = ClassTag(classOf[LogProcessorMessage])
  override def schema: StructType = null 
}

这篇关于Spark 2.x Scala 2.1.1自定义编码器类类型不匹配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IT屋！