SparkSQL stuff

df.scala

import org.apache.spark.SparkContext
import org.apache.spark.sql.SQLContext

val a = Array(1, 2, 3, 4)
val b = Array("one", "two", "three", "four")
val sc = SparkContext.getOrCreate()
val sql = SQLContext.getOrCreate(sc)
val df = sc.parallelize(a.zip(b)).toDF("number", "word")

# Collects the dataframe
df.collect()

# Convert back to a RDD[(Int, String)]
df.map { case Row(i: Int, w: String) => (i, w)}


df.schema