KSQL 可以使用模式中提供的默认值填充流吗？

Question

I have a KSQL stream that I'd like to populate with the default values specified in the schema. Other than manually specifying them again with coalesce statements, is there a way to do this?

The previously answered question is similar to the issue I am facing, but it doesn't address the main point of using the default values already specified in the schema:

Create KSQL stream with default values for a column?

I did the following (based on the documentation provided by Confluent: https://docs.confluent.io/platform/current/schema-registry/serdes-develop/serdes-avro.html#schema-references-in-avro ):

1. Created a topic t1-a with a schema

kafka-avro-console-producer --bootstrap-server localhost:9092 --property schema.registry.url=http://localhost:8081 --topic t1-a \
--property value.schema='{"type":"record","name":"myrecord","fields":[{"name":"name","type":"string","default":"no-name"}]}

2. Set the compatibility to FULL on the subject (using the schema-registry REST API)
3. Generated records to the topic using the CLI tool

{"name":"john"}
{"name":"doe"}

4. Updated the schema using the Schema-registry

kafka-avro-console-producer --bootstrap-server localhost:9092 --property schema.registry.url=http://localhost:8081 --topic t1-a \
--property value.schema='{"type":"record","name":"myrecord","fields":[{"name":"name","type":"string", "default":"no-name"}, {"name":"age","type":"string", "default":"ageless-wonder"}]}

5. Generated records to the topic using the CLI tool:

{"name":"jack", "age":"100"}
{"name":"jill", "age":"101"}

6. started ksql cli and created a stream

CREATE STREAM t1_a WITH (KAFKA_TOPIC='t1-a',VALUE_FORMAT='AVRO');

7. Queried Records:

SELECT * FROM t1_a;

Now I get the records, but the value in Age for John and Doe are listed as null (instead of the default value "ageless-wonder" specified in the schema):

NAME    AGE
john    null
doe     null
jack    100
jill    101

I understand that I can coalesce the values to default in the stream definition, but is there a way to have that field populate based on the schema already provided?

Answer 1

Ksql uses the schema ID that was originally sent with the data, rather than the latest, to deserialize the record and build each row. You'd need to define your first schema with a default age, then send without.