scala - Scala 在 if else 语句中赋值

Question

对scala来说很新。

我有一个 if else 语句，它基于条件为变量赋值。

import org.apache.spark.ml.{Pipeline, PipelineModel, PipelineStage}
modelType = "rf" 

def train(trainingData: DataFrame): PipelineModel = {

        val xgb = new XGBoostClassifier()
          .setLabelCol("label")
          .setFeaturesCol("features")
          .setObjective("binary:logistic")
          .setEvalMetric("auc")  

        val rf = new RandomForestClassifier()
          .setLabelCol("label")
          .setFeaturesCol("features")
          .setImpurity("gini")

         if (modelType == "rf") {
             val pipeline = new Pipeline()
              .setStages(Array[PipelineStage](rf))
             } else if (modelType == "xgb") {
            val pipeline = new Pipeline()
              .setStages(Array[PipelineStage](rf))
             }


        pipeline.fit(trainingData)
        }

但我看到这个错误：pipeline not found。这是否意味着未分配管道变量？我该如何重写这个？

标签： scalaapache-sparkapache-spark-sqlapache-spark-mllib

Answer 1

您将 if 设置为变量

  val xgb = new XGBoostClassifier()
              .setLabelCol("label")
              .setFeaturesCol("features")
              .setObjective("binary:logistic")
              .setEvalMetric("auc")  

            val rf = new RandomForestClassifier()
              .setLabelCol("label")
              .setFeaturesCol("features")
              .setImpurity("gini")

           val pipeline =  if (modelType == "rf") {
                  new Pipeline()
                  .setStages(Array[PipelineStage](rf))
                 } else if (modelType == "xgb") {
                new Pipeline()
                  .setStages(Array[PipelineStage](rf))
                 }
                  }

            pipeline.fit(trainingData)
            }