spark 2.1 TaskResult

来源:互联网 发布:知天命尽人事的网名 编辑:程序博客网 时间:2024/06/06 07:35

TaskResult

// Task result. Also contains updates to accumulator variables.private[spark] sealed trait TaskResult[T]

IndirectTaskResult

/** A reference to a DirectTaskResult that has been stored in the worker's BlockManager. */private[spark] case class IndirectTaskResult[T](blockId: BlockId, size: Int)  extends TaskResult[T] with Serializable

DirectTaskResult

/** A TaskResult that contains the task's return value and accumulator updates. */private[spark] class DirectTaskResult[T](    var valueBytes: ByteBuffer,    var accumUpdates: Seq[AccumulatorV2[_, _]])  extends TaskResult[T] with Externalizable {  private var valueObjectDeserialized = false  private var valueObject: T = _  def this() = this(null.asInstanceOf[ByteBuffer], null)  override def writeExternal(out: ObjectOutput): Unit = Utils.tryOrIOException {    out.writeInt(valueBytes.remaining)    Utils.writeByteBuffer(valueBytes, out)    out.writeInt(accumUpdates.size)    accumUpdates.foreach(out.writeObject)  }  override def readExternal(in: ObjectInput): Unit = Utils.tryOrIOException {    val blen = in.readInt()    val byteVal = new Array[Byte](blen)    in.readFully(byteVal)    valueBytes = ByteBuffer.wrap(byteVal)    val numUpdates = in.readInt    if (numUpdates == 0) {      accumUpdates = Seq()    } else {      val _accumUpdates = new ArrayBuffer[AccumulatorV2[_, _]]      for (i <- 0 until numUpdates) {        _accumUpdates += in.readObject.asInstanceOf[AccumulatorV2[_, _]]      }      accumUpdates = _accumUpdates    }    valueObjectDeserialized = false  }  /**   * When `value()` is called at the first time, it needs to deserialize `valueObject` from   * `valueBytes`. It may cost dozens of seconds for a large instance. So when calling `value` at   * the first time, the caller should avoid to block other threads.   *   * After the first time, `value()` is trivial and just returns the deserialized `valueObject`.   */  def value(resultSer: SerializerInstance = null): T = {    if (valueObjectDeserialized) {      valueObject    } else {      // This should not run when holding a lock because it may cost dozens of seconds for a large      // value      val ser = if (resultSer == null) SparkEnv.get.serializer.newInstance() else resultSer      valueObject = ser.deserialize(valueBytes)      valueObjectDeserialized = true      valueObject    }  }}