Spark Shell — spark-shell shell script

Spark shell is an interactive environment where you can learn how to make the most out of Apache Spark quickly and conveniently.

Spark shell is particularly helpful for fast interactive prototyping.

Under the covers, Spark shell is a standalone Spark application written in Scala that offers environment with auto-completion (using TAB key) where you can run ad-hoc queries and get familiar with the features of Spark (that help you in developing your own standalone Spark applications). It is a very convenient tool to explore the many things available in Spark with immediate feedback. It is one of the many reasons why Spark is so helpful for tasks to process datasets of any size.

There are variants of Spark shell for different languages: spark-shell for Scala, pyspark for Python and sparkR for R.

This document (and the book in general) uses spark-shell for Scala only.

You can start Spark shell using spark-shell script.

$ ./bin/spark-shell

spark-shell is an extension of Scala REPL with automatic instantiation of SparkSession as spark (and SparkContext as sc).

scala> :type spark

// Learn the current version of Spark in use
scala> spark.version
res0: String = 2.1.0-SNAPSHOT

spark-shell also imports Scala SQL’s implicits and sql method.

scala> :imports
 1) import spark.implicits._       (59 terms, 38 are implicit)
 2) import spark.sql               (1 terms)

When you execute spark-shell you actually execute Spark submit as follows:

org.apache.spark.deploy.SparkSubmit --class org.apache.spark.repl.Main --name Spark shell spark-shell

Set SPARK_PRINT_LAUNCH_COMMAND to see the entire command to be executed. Refer to Print Launch Command of Spark Scripts.

Using Spark shell

You start Spark shell using spark-shell script (available in bin directory).

$ ./bin/spark-shell
Setting default log level to "WARN".
To adjust logging level use sc.setLogLevel(newLevel). For SparkR, use setLogLevel(newLevel).
WARN NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
WARN ObjectStore: Failed to get database global_temp, returning NoSuchObjectException
Spark context Web UI available at
Spark context available as 'sc' (master = local[*], app id = local-1477858597347).
Spark session available as 'spark'.
Welcome to
      ____              __
     / __/__  ___ _____/ /__
    _\ \/ _ \/ _ `/ __/  '_/
   /___/ .__/\_,_/_/ /_/\_\   version 2.1.0-SNAPSHOT

Using Scala version 2.11.8 (Java HotSpot(TM) 64-Bit Server VM, Java 1.8.0_112)
Type in expressions to have them evaluated.
Type :help for more information.


Spark shell creates an instance of SparkSession under the name spark for you (so you don’t have to know the details how to do it yourself on day 1).

scala> :type spark

Besides, there is also sc value created which is an instance of SparkContext.

scala> :type sc

To close Spark shell, you press Ctrl+D or type in :q (or any subset of :quit).

scala> :q


Table 1. Spark Properties
Spark Property Default Value Description



Used in spark-shell to create REPL ClassLoader to load new classes defined in the Scala REPL as a user types code.

Enable INFO logging level for org.apache.spark.executor.Executor logger to have the value printed out to the logs:

INFO Using REPL class URI: [classUri]