Hi,
I would like to propose that we have an optional schema property on the Avro 
format. This would be a String of the Avro schema and should be supplied as the 
correct schema to be supplied in the format. This is required because the 
schema that is inferred from the table definition is not always accurate; for 
example with nested objects with child objects that have different nullability 
to the parent.

Confluent Avro schema format has this option already.

This will be used as the writeSchema [1] for Avro format.

I am happy to implement this,

Kind regards, David.

[1] 
https://github.com/apache/flink/blob/271266a9cd5e4b5d3d564c956cfcda76cb60fa36/flink-formats/flink-avro/src/main/java/org/apache/flink/formats/avro/RegistryAvroDeserializationSchema.java#L97

Unless otherwise stated above:

IBM United Kingdom Limited
Registered in England and Wales with number 741598
Registered office: Building C, IBM Hursley Office, Hursley Park Road, 
Winchester, Hampshire SO21 2JN

Reply via email to