value toDF is not a member of org.apache.spark.rdd.RDD

Question

I've read about this issue in other SO posts and I still don't know what I'm doing wrong. In principle, adding these two lines:

val sqlContext = new org.apache.spark.sql.SQLContext(sc)
import sqlContext.implicits._

should have done the trick but the error persists

This my build.sbt:

name := "PickACustomer"

version := "1.0"

scalaVersion := "2.11.7"


libraryDependencies ++= Seq("com.databricks" %% "spark-avro" % "2.0.1",
"org.apache.spark" %% "spark-sql" % "1.6.0",
"org.apache.spark" %% "spark-core" % "1.6.0")

and my scala code is:

import scala.collection.mutable.Map
import scala.collection.immutable.Vector

import org.apache.spark.rdd.RDD
import org.apache.spark.SparkContext
import org.apache.spark.SparkContext._
import org.apache.spark.SparkConf
import org.apache.spark.sql._


    object Foo{

    def reshuffle_rdd(rawText: RDD[String]): RDD[Map[String, (Vector[(Double, Double, String)], Map[String, Double])]]  = {...}

    def do_prediction(shuffled:RDD[Map[String, (Vector[(Double, Double, String)], Map[String, Double])]], prediction:(Vector[(Double, Double, String)] => Map[String, Double]) ) : RDD[Map[String, Double]] = {...}

    def get_match_rate_from_results(results : RDD[Map[String, Double]]) : Map[String, Double]  = {...}


    def retrieve_duid(element: Map[String,(Vector[(Double, Double, String)], Map[String,Double])]): Double = {...}




    def main(args: Array[String]){
        val conf = new SparkConf().setAppName(this.getClass.getSimpleName)
        if (!conf.getOption("spark.master").isDefined) conf.setMaster("local")

        val sc = new SparkContext(conf)

        //This should do the trick
        val sqlContext = new org.apache.spark.sql.SQLContext(sc)
        import sqlContext.implicits._

        val PATH_FILE = "/mnt/fast_export_file_clean.csv"
        val rawText = sc.textFile(PATH_FILE)
        val shuffled = reshuffle_rdd(rawText)

        // PREDICT AS A FUNCTION OF THE LAST SEEN UID
        val results = do_prediction(shuffled.filter(x => retrieve_duid(x) > 1) , predict_as_last_uid)
        results.cache()

        case class Summary(ismatch: Double, t_to_last:Double, nflips:Double,d_uid: Double, truth:Double, guess:Double)

        val summary = results.map(x => Summary(x("match"), x("t_to_last"), x("nflips"), x("d_uid"), x("truth"), x("guess")))


        //PROBLEMATIC LINE
        val sum_df = summary.toDF()

    }
    }

I always get:

value toDF is not a member of org.apache.spark.rdd.RDD[Summary]

Bit lost now. Any ideas?

Answer 1

Move your case class outside of main:

object Foo {

  case class Summary(ismatch: Double, t_to_last:Double, nflips:Double,d_uid: Double, truth:Double, guess:Double)

  def main(args: Array[String]){
    ...
  }

}

Something about the scoping of it is preventing Spark from being able to handle the automatic derivation of the schema for Summary. FYI I actually got a different error from sbt:

No TypeTag available for Summary