When case classes cannot be created in advance. e.g. schema is being read from a JSON file, we cannot use reflection method i.e. cannot use rdd.map(x=> case_class(x._1,x._2)).toDF()

rdd-to-DF-NOreflection.scala

//Programmatically Specifying the Schema


import org.apache.spark.{SparkContext,SparkConf}
import org.apache.spark.sql.hive.HiveContext

//initializations
val conf = new SparkConf().setAppName("xx").setMaster("local[2]")
val sc = new SparkContext(conf)
val hiveContext = new HiveContext(sc)
//

//create DF from RDD
import org.apache.spark.sql.Row
import org.apache.spark.sql.types.{StructType, StructField, IntegerType}

val schema = StructType( StructField("column", IntegerType, false) :: Nil)
//pyspark: 
  //from pyspark.sql.types import IntegerType,StructType,StructField
  //qlContext.createDataFrame(rdd,StructType([StructField("col1",IntegerType(),False),StructField("col2",IntegerType(),False)]))

val df = hiveContext.createDataFrame(rdd.map(x=> Row(x)),schema)
df.show