Apache Flink Training Excercises

Clone this repo:
  1. e598712 [hotfix] updated READMEs by Tomas · 2 months ago master
  2. d3914a2 [FLINK-20597] update for Flink 1.12 by Svend Vanderveken · 3 months ago release-1.12
  3. 40fdc82 [hotfix] output examples in HourlyTips discussion should match actual output (#14) by David Anderson · 7 months ago
  4. ea4a66e [FLINK-18630] Redo the LongRideAlerts solutions to cover a corner case (#13) by David Anderson · 8 months ago release-1.11
  5. 4f95a44 [FLINK-18482] Replace file-based data sources with generators by David Anderson · 10 months ago

Apache Flink Training Exercises

Exercises that go along with the training content in the documentation.

Table of Contents

Setup your Development Environment

  1. Software requirements
  2. Clone and build the flink-training project
  3. Import the flink-training project into your IDE

Using the Taxi Data Streams

  1. Schema of Taxi Ride Events
  2. Generating Taxi Ride Data Streams in a Flink program

How to do the Labs

  1. Learn about the data
  2. Modify ExerciseBase
  3. Run and debug Flink programs in your IDE
  4. Exercises, Tests, and Solutions



Setup your Development Environment

The following instructions guide you through the process of setting up a development environment for the purpose of developing, debugging, and executing solutions to the Flink developer training exercises and examples.

Software requirements

Flink supports Linux, OS X, and Windows as development environments for Flink programs and local execution. The following software is required for a Flink development setup and should be installed on your system:

  • a JDK for Java 8 or Java 11 (a JRE is not sufficient; other versions of Java are not supported)
  • Git
  • an IDE for Java (and/or Scala) development with Gradle support. We recommend IntelliJ, but Eclipse or Visual Studio Code can also be used so long as you stick to Java. For Scala you will need to use IntelliJ (and its Scala plugin).

:information_source: Note for Windows users: The examples of shell commands provided in the training instructions are for UNIX systems. To make things easier, you may find it worthwhile to setup cygwin or WSL. For developing Flink jobs, Windows works reasonably well: you can run a Flink cluster on a single machine, submit jobs, run the webUI, and execute jobs in the IDE.

Clone and build the flink-training project

This flink-training project contains exercises, tests, and reference solutions for the programming exercises. Clone the flink-training project from Github and build it.

:information_source: Repository Layout: This repository has several branches set up pointing to different Apache Flink versions, similarly to the apache/flink repository with:

  • a release branch for each minor version of Apache Flink, e.g. release-1.10, and
  • a master branch that points to the current Flink release (not flink:master!)

If you want to work on a version other than the current Flink release, make sure to check out the appropriate branch.

git clone https://github.com/apache/flink-training.git
cd flink-training
./gradlew test shadowJar

If you haven’t done this before, at this point you’ll end up downloading all of the dependencies for this Flink training project. This usually takes a few minutes, depending on the speed of your internet connection.

If all of the tests pass and the build is successful, you are off to a good start.

If you are in China, we recommend configuring the maven repository to use a mirror. You can do this by uncommenting the appropriate line in our build.gradle like this:

    repositories {
        // for access from China, you may need to uncomment this line
        maven { url 'http://maven.aliyun.com/nexus/content/groups/public/' }
        maven {
            url "https://repository.apache.org/content/repositories/snapshots/"
            mavenContent {

Import the flink-training project into your IDE

The project needs to be imported as a gradle project into your IDE.

Once that’s done you should be able to open RideCleansingTest and successfully run this test.

:information_source: Note for Scala users: You will need to use IntelliJ with the JetBrains Scala plugin, and you will need to add a Scala 2.12 SDK to the Global Libraries section of the Project Structure. IntelliJ will ask you for the latter when you open a Scala file.

Using the Taxi Data Streams

These exercises use data generators that produce simulated event streams inspired by those shared by the New York City Taxi & Limousine Commission in their public data set about taxi rides in New York City.

Schemas of Taxi Ride and Taxi Fare Events

Our taxi data set contains information about individual taxi rides in New York City. Each ride is represented by two events: a trip start, and a trip end event. Each event consists of eleven fields:

rideId         : Long      // a unique id for each ride
taxiId         : Long      // a unique id for each taxi
driverId       : Long      // a unique id for each driver
isStart        : Boolean   // TRUE for ride start events, FALSE for ride end events
startTime      : Instant   // the start time of a ride
endTime        : Instant   // the end time of a ride,
                           //   "1970-01-01 00:00:00" for start events
startLon       : Float     // the longitude of the ride start location
startLat       : Float     // the latitude of the ride start location
endLon         : Float     // the longitude of the ride end location
endLat         : Float     // the latitude of the ride end location
passengerCnt   : Short     // number of passengers on the ride

There is also a related data set containing fare data about those same rides, with these fields:

rideId         : Long      // a unique id for each ride
taxiId         : Long      // a unique id for each taxi
driverId       : Long      // a unique id for each driver
startTime      : Instant   // the start time of a ride
paymentType    : String    // CASH or CARD
tip            : Float     // tip for this ride
tolls          : Float     // tolls for this ride
totalFare      : Float     // total fare collected

How to do the Labs

In the hands-on sessions you will implement Flink programs using various Flink APIs.

The following steps guide you through the process of using the provided data streams, implementing your first Flink streaming program, and executing your program in your IDE.

We assume you have setup your development environment according to our setup guide above.

Learn about the data

The initial set of exercises are all based on data streams of events about taxi rides and taxi fares. These streams are produced by source functions which reads data from input files. Please read the instructions above to learn how to use them.

Run and debug Flink programs in your IDE

Flink programs can be executed and debugged from within an IDE. This significantly eases the development process and provides an experience similar to working on any other Java (or Scala) application.

Starting a Flink program in your IDE is as easy as running its main() method. Under the hood, the execution environment will start a local Flink instance within the same process. Hence it is also possible to put breakpoints in your code and debug it.

Assuming you have an IDE with this flink-training project imported, you can run (or debug) a simple streaming job as follows:

  • Open the org.apache.flink.training.examples.ridecount.RideCountExample class in your IDE
  • Run (or debug) the main() method of the RideCountExample class using your IDE.

Exercises, Tests, and Solutions

Each of these exercises includes an ...Exercise class with most of the necessary boilerplate code for getting started, as well as a JUnit Test class (...Test) with a few tests for your implementation, and a ...Solution class with a complete solution.

:information_source: Note: As long as your ...Exercise class is throwing a MissingSolutionException, the provided JUnit test classes will ignore that failure and verify the correctness of the solution implementation instead.

There are Java and Scala versions of all the exercise, test, and solution classes.

Now you are ready to begin with the first exercise in our Labs.


The code in this repository is licensed under the Apache Software License 2.