Parsing and Querying CSVs With Apache Spark

In this post, I will be sharing how to parse and query CSV data using Apache Spark. Querying CSV data is very easy using the Spark CSV library. For that, we will be using SQLContext. With SQLContext, we can query the data like we do in any database language. We can perform all the operation on data like SELECT and also write the data into a new file.

Alright, enough talking now, let’s see how we can do it!

We will set up an SBT project first. After setting up an SBT project, we will start by adding required dependencies into build.sbt.

For this project, we will require 3 major dependencies:

libraryDependencies += "org.apache.spark" %% "spark-core" % "1.6.1"
//for spark 1.6 
libraryDependencies += "org.apache.spark" %% "spark-sql" % "1.6.1"
// for spark sqlcontext
libraryDependencies += "com.databricks" % "spark-csv_2.11" % "1.5.0"
// this is the show stealer,this is the dependency 
required to parse and query csv data

Alright, now we are done setting up the project and adding dependencies. We will be writing the code in steps so that it will be easier for you to understand each and every line. 

val sparkConf = new SparkConf().setAppName("simpleReading").
setMaster("local[2]") //set spark configuration

val sparkContext = new SparkContext(sparkConf) // make spark context

val sqlContext = new SQLContext(sparkContext) // make sql context

val df = sqlContext.read.
    format("com.databricks.spark.csv").
    option("header","true").
    option("inferSchema","true").
load("employee.csv") //load data from a file

val selectedCity = df.select("city")
  selectedCity.write.save("employee.csv") 
//select a particular column which is "city" from the file and save the selected data into a  new csv

val selectNameAndAge = df.select("name","age") //select 
particular columns
  selectMake.show() //show the selected columns

val tempTable = df.registerTempTable("my_table") 
//makes a temporary table my_table
  val usingSQL = sqlContext.sql("select * from my_table") 
//select all the csv file's data in temp table
usingSQL.show() //show the temporary table using show function 

Final Code

import org.apache.spark.sql.SQLContext
import org.apache.spark.{SparkConf, SparkContext}

/**
 * Created by deepak on 7/1/17.
 */
object SparkTest extends App {
  val sparkConf = new SparkConf().setAppName("simpleReading").
setMaster("local[2]")
  //set spark configuration
  val sparkContext = new SparkContext(sparkConf)
  // make spark context
  val sqlContext = new SQLContext(sparkContext) // make sql context

  val df = sqlContext.read.
    format("com.databricks.spark.csv").
    option("header", "true").
    option("inferSchema", "true").load("employee.csv") 
//load data from a file

  val selectedCity = df.select("city")
  selectedCity.write.save("employee.csv")
  //save the data in new csv

  val selectMake = df.select("name", "age") //select particular column
  selectMake.show()
  //show make column

  val tempTable = df.registerTempTable("my_table")
  //makes a temporary table
  val usingSQL = sqlContext
    .sql("select * from my_table") 
//show all the csv file's data in temp table
  usingSQL.show()
}

You can find the mini project with all the code at the link Click here. If you find any challenge, Do let me know in the comments. If you enjoyed this post, I’d be very grateful if you’d help it spread.

Keep smiling, Keep coding!