import session.implicits._ doesn't work

[raviteja-ms]

Spark Version: 2.0.2
Scala Version: 2.11.8

When I tried a simple example to import spark.implicits._ I get the following error while sbt build:

My code:

import io.hydrosphere.mist.lib.{MistJob, SQLSupport}
import org.apache.spark.sql._

object SimpleContext extends MistJob with SQLSupport{
  /** Contains implementation of spark job with ordinary [[org.apache.spark.SparkContext]]
    * Abstract method must be overridden
    *
    * based on https://github.com/Hydrospheredata/mist/blob/master/examples/src/main/scala/SimpleContext.scala
    *
    * @param digits user parameter digits
    * @return result of the job
    */
 import session.implicits._
 def execute(digits:Seq[Int]): Map[String, Any] = {
        val mydf = digits.toSeq.toDF("number")
        Map("result" -> mydf.map(x => x.getInt(0) * 3).collect())
  }
}

Error:

[info] Loading project definition from /home/user/mist/apache-spark-restAPI-example/project
[info] Set current project to sparkMist (in build file:/home/user/mist/apache-spark-restAPI-example/)
[info] Compiling 1 Scala source to /home/user/mist/apache-spark-restAPI-example/target/scala-2.11/classes...
[error] /home/user/mist/apache-spark-restAPI-example/src/main/scala/SimpleContext.scala:13: stable identifier required, but SimpleContext.this.session.implicits found.
[error]  import session.implicits._
[error]                 ^
[error] one error found
[error] (compile:compileIncremental) Compilation failed

Please help me how I can import spark.implicits._

Thanks..!!

[dos65]

Currently inside MistJob session is defined as a function, but should be a value. We will fix it later.
As a workaround I can suggest to assign session to some variable and then call implicits import from it.

import io.hydrosphere.mist.lib.{MistJob, SQLSupport}
import org.apache.spark.sql._

object SimpleContext extends MistJob with SQLSupport {

   val s = session
   import s.implicits._  

  def execute(digits:Seq[Int]): Map[String, Any] = {
      val mydf = digits.toSeq.toDF("number")
      Map("result" -> mydf.map(x => x.getInt(0) * 3).collect())
  }
}

[raviteja-ms]

Thanks, this worked...!!
But, please mention this documentation or in any example somewhere. I was confused and other new users will be too.