Apache Kafka 1.0.0 Streams API Multiple Layered Group

Question

How to use .groupby with few restrictions in the Kafka Streams API. Same as Java Streams API Example below

public void twoLevelGrouping(List<Person> persons) {
     final Map<String, Map<String, List<Person>>> personsByCountryAndCity = persons.stream().collect(
         groupingBy(Person::getCountry,
            groupingBy(Person::getCity)
        )
    );
    System.out.println("Persons living in London: " + personsByCountryAndCity.get("UK").get("London").size());
}

Answer 1

You specify a combination key by putting all the attributes / fields that you want to group into a key.

KTable table = stream.selectKey((k, v,) -> k::getCountry + "-" + k::getCity)
                     .groupByKey()
                     .aggregate(...); // or maybe .reduce()

I just assumed that the country and the city are both String. You are using Interactive Queries to query the repository using

store.get("UK-London");

https://docs.confluent.io/current/streams/developer-guide/interactive-queries.html