kafka-strams leftJoin produce NPE - apache-kafka-streams

Question

I have the next values in topics
// photos
1, {"id": 1, user_id: 1, "url": "http://example.com"}
2, {"id": 2, user_id: 2, "url": "http://example1.com"}
3, {"id": 3, user_id: 1, "url": "http://example2.com"}
// users
1, {"id": 1, "name": "user1"}
2, {"id": 2, "name": "user2"}
I want to get info: [photo_id, photo_url, user_id, user_name]
I implement Result class for it
public class Result {
public int photo_id;
public String photo_url;
public int user_id;
public String user_name;
public static Result from(Photo photo, User user) {
Result r = new Result();
r.photo_id = photo.id;
r.photo_url = r.url;
r.user_id = user.id;
r.user_name = user.name;
return r;
}
}
And my stream implementation:
final KStream<Integer, Photo) photo_by_user = ...;
final KStream<Integer, User) users = ...;
users.leftJoin(photo_by_user, new ValueJoiner<User, Photo, Result> {
public Result apply(User user, Photo photo) {
return Result.from(user, photo);
}
}, JoinWindows.of(1L))
But when I run this code, I get:
Exception in thread "example-StreamThread-1" java.lang.NullPointerException
at myapps.util.Result.create(Result.java:15)
at myapps.Example$1.apply(Example.java:56)
at myapps.Example$1.apply(Example.java:53)
at org.apache.kafka.streams.kstream.internals.KStreamKStreamJoin$KStreamKStreamJoinProcessor.process(KStreamKStreamJoin.java:87)
at org.apache.kafka.streams.processor.internals.ProcessorNode$1.run(ProcessorNode.java:46)
Because User have a value, but Photo is null
But I do not understand why? And how to avoid it.

Answer 1

with leftJoin you can have null Photo values inside ValueJoiner apply method.
for each input record of users stream that does not satisfy the join predicate the provided ValueJoiner will be called with a null value for the photo_by_user stream.
inside Result.from() method you need to check that Photo instance is non null, and only after that get photo's fields id and url.
Also you have left join with JoinWindows.of(1L), where 1 is millisecond.
pay attention that two records are only joined if their timestamps are close to each other as defined by specified JoinWindows. potentially in your case two records were created not exactly at the same time, so try to increase value e.g. 10000L to test your joining logic.