Real time data streaming workshop

Practical Exercises of Kafka Streams API

Exercise 1

/** *
 * 1. Define properties for application
 * In createTopology method:
 *  a. Create StreamsBuilder instance
 *  b. Create source KStream out of input topic 'ex1-input'
 *  c. Sink kstream to output topic 'ex1-stream-output'
 *  d. Create sourceKTable out of input topic 'ex1-input'
 *  e. Sink ktable to output topic 'ex1-table-output'
 * 2. Create KafkaStreams instance
 * 3. Start kafka streams
 */

Exercise 2

/***
 * Unified orders system produce data to 1 input topic for legacy and new systems
 * Message's key is an indicator of legacy or new system [legacy|new]
 * Messages start with product name (bambi, godzilla, optimus) then is separator':' batch of quantity
 * numbers separated with commas (number can be negative for number of returned pieces) e.q {productName}:1,1,2,-3
 * Filter invalid products as our analysis needs only bambi, godzilla and optimus to be investigated
 * as a result we need create 2 output topics ex2-legacy-out and ex2-new-output each of them
 * should have sum for each of desired products
 *
 * You can assume no malformed records
 *
 */

Exercise 3

/***
 *
 * Create global table of user data and stream of orders then join orders data
 * with inner and left join so that message with remain having key of order
 * and value in format of "Order=orderValue, UserInfo=[userInfoValue]"
 * take into account 'null' values in case of left join
 *
 */

Exercise 4

/***
 * You need to join 2 topics ex4-doctors and ex4-clinics both are json/POJO serialized
 * Serdes and Serializers have already been prepared for you
 * For both doctors and clinics its primary key is used as message key
 * This time key for join is inside of Clinic pojo (dockorId)
 * You will have to figure out how to use Kafka Streams API methods to create one-to-many join
 * Then you will have to flatten that one-to-many structure
 * to get final output topic ex4-output in format:
 * long value key (doctor id)
 * string value message {doctor.id}-{doctor.specialityCode}-{clinic.id}
 *
 */

Exercise 5

/***
 * This time run docker-compose down to shut down previous run
 * and start docker-compose --file docker-compose-ex5.yaml up -d
 *
 * There are 3 topics streamed via kafka connect
 * ex5-doctors, ex5-diseases, ex5-prescriptions
 * Each of them has key in json format {"id": 1} and json value
 * For both ID (ID is name of model for key) and each value format
 * models/serializers/deserializers/serdes have already been prepared for you
 * you will have take a look on data and join/map/aggregate however you want
 * to obtain at the end ex5-output topic where key will be long value diseaseId from Disease pojo
 * and value will be json containing doctorId and prescripted medicineName for that disease
 *
 * Take a note that you will have to introduce new data types, serializers, deserializers and serdes
 * that is easy you can base on implementation done in infra
 *
 * Intentionally there is no test for that exercise
 *
 */