Hey, I have a question regarding Avro Types and schema evolution. According to the docs the schema resolution is compatible with the Avro docs [1].
But I have done some testing. For example, I have created a record, written it to Kafka, and then changed the order the fields in schema and tried to read the data with changed schema using AvroDeserializationSchema from Flink, it was failing with IndexOfBounds during deserialization, but according to the docs fields are resolved by name during deserialization and changed ordering should not really be a problem. I have also found some other inconsistencies like when the writer is missing the field, but it is optional for reader, which means that it has default value. According to Avro docs this should be resolved, but I only managed to resolve this of the missing field is at the end, not if it's in the middle. Is there anything that I don't understand here, or is something wrong ?? Best Regards, Dom. [1]http://avro.apache.org/docs/current/spec.html#Schema+Resolution
