spark create dataset from scala list

In this short article I will show how to create dataframe/dataset in spark sql.

In scala we can use the tuple objects to simulate the row structure if the number of column is less than or equal to 22 . Lets say in our example we want to create a dataframe/dataset of 4 rows , so we will be using Tuple4 class. Below is the example of the same


import org.apache.spark.sql.{DataFrame, SparkSession}
import scala.collection.mutable.ListBuffer

class SparkDataSetFromList {

def getSampleDataFrameFromList(sparkSession: SparkSession): DataFrame = {

import sparkSession.implicits._

var sequenceOfOverview = ListBuffer[(String, String, String, Integer)]()
sequenceOfOverview += Tuple4("Apollo", "1", "20200901", 1)
sequenceOfOverview += Tuple4("Apollo", "2", "20200901", 0)
sequenceOfOverview += Tuple4("Apollo", "3", "20200901", 1)
sequenceOfOverview += Tuple4("Apollo", "4", "20200901", 0)

sequenceOfOverview += Tuple4("Apollo", "1", "20200902", 1)
sequenceOfOverview += Tuple4("Apollo", "2", "20200902", 0)
sequenceOfOverview += Tuple4("Apollo", "3", "20200902", 1)
sequenceOfOverview += Tuple4("Apollo", "4", "20200902", 1)

sequenceOfOverview += Tuple4("Apollo", "1", "20200903", 0)
sequenceOfOverview += Tuple4("Apollo", "2", "20200903", 0)
sequenceOfOverview += Tuple4("Apollo", "3", "20200903", 0)
sequenceOfOverview += Tuple4("Apollo", "4", "20200903", 1)

sequenceOfOverview += Tuple4("Apollo", "1", "20200904", 0)
sequenceOfOverview += Tuple4("Apollo", "2", "20200904", 0)
sequenceOfOverview += Tuple4("Apollo", "3", "20200904", 1)
sequenceOfOverview += Tuple4("Apollo", "4", "20200904", 1)

val df1 =
sequenceOfOverview.toDF("Hospital", "AccountNumber", "date", "Visit")
df1
}

}

Let`s print the schema and values to examine the same


import org.apache.spark.sql.SparkSession

object Gateway extends App {

val sparkExcel = new SparkExcel()
lazy val sparkSession: SparkSession = SparkSession
.builder()
.master("local[*]")
.getOrCreate()
val dataframe = sparkExcel.getSampleDataFrame(sparkSession)
dataframe.printSchema()
dataset.show
}

Below is the result

root
|-- Hospital: string (nullable = true)
|-- AccountNumber: string (nullable = true)
|-- date: string (nullable = true)
|-- Visit: integer (nullable = true)

 

+--------+-------------+--------+-----+
|Hospital|AccountNumber| date|Visit|
+--------+-------------+--------+-----+
| Apollo| 1|20200901| 1|
| Apollo| 2|20200901| 0|
| Apollo| 3|20200901| 1|
| Apollo| 4|20200901| 0|
| Apollo| 1|20200902| 1|
| Apollo| 2|20200902| 0|
| Apollo| 3|20200902| 1|
| Apollo| 4|20200902| 1|
| Apollo| 1|20200903| 0|
| Apollo| 2|20200903| 0|
| Apollo| 3|20200903| 0|
| Apollo| 4|20200903| 1|
| Apollo| 1|20200904| 0|
| Apollo| 2|20200904| 0|
| Apollo| 3|20200904| 1|
| Apollo| 4|20200904| 1|
+--------+-------------+--------+-----+

That’s a brief on how we can create dataframe/dataset from scala list in spark sql.

Leave a Reply

Your email address will not be published. Required fields are marked *