CODEWITHSUNDEEP
joinstreamapache-kafkaksqldb
Tim Coombe
Tim Coombe

Reputation: 33

KSQL Join more than two streams

Is it possible to join more than two streams/tables in KSQL?

Example:

I have three streams:

CREATE STREAM StreamA (id BIGINT, message VARCHAR) WITH 
(KAFKA_TOPIC='TopicA', VALUE_FORMAT='DELIMITED');
CREATE STREAM StreamB (id BIGINT, aid BIGINT, message VARCHAR) WITH . 
(KAFKA_TOPIC='TopicB', VALUE_FORMAT='DELIMITED');
CREATE STREAM StreamC (id BIGINT, bid BIGINT, message VARCHAR) WITH 
(KAFKA_TOPIC='TopicC', VALUE_FORMAT='DELIMITED');

I try to create another stream by joining those three streams:

CREATE STREAM ABCStream AS SELECT * FROM StreamA a JOIN 
StreamB b ON b.aid = a.id JOIN StreamC c WITHIN 1 HOURS ON 
c.bid = b.id;

I get the following exception:

mismatched input 'JOIN' expecting ';'  
Caused by: org.antlr.v4.runtime.InputMismatchException

Upvotes: 3

Views: 2765

Answers (1)

Robin Moffatt
Robin Moffatt

Reputation: 32090

No, you can only join two per query in KSQL up to v5.0. You'd need to daisy-chain your queries, something like this:

Intermediate Stream: 

CREATE STREAM ABStream AS \
   SELECT * \
     FROM StreamA a \
     JOIN StreamB b \
          ON b.aid = a.id;

Multi- join stream

CREATE STREAM ABCStream AS \
   SELECT * \
     FROM ABStream AB \
     JOIN StreamC c \
          WITHIN 1 HOURS \
          ON c.bid = AB.b_id;

Upvotes: 5

Related Questions