Maximilian Michels
Maximilian Michels
I think there are different concerns here: (a) data integrity and (b) data semantics. (a) The value of a field should always return its correct value. That's why the Flink...
>My intuitive understanding of the mentioned problem in Iceberg is why not fix the nullability of the field? In Iceberg, users can directly provide a stream of RowData, for which...
> For the 1st part, my question is since the nullability info provided by users are not reliable, why not enforce the use of nullable type when use `RowData#createFieldGetter`? Am...
@twalthr Do you mind having a look?
@twalthr Rebased. It may still be worthwhile to merge the tests, to prevent regressions regarding null values.
@davidradl Thanks for chiming in! IMHO the title still reflects the change, but I can adjust the title to focus more on the tests, e.g.: "[FLINK-37245] Add tests for FieldGetters...
@davidradl @twalthr Could you have another look?
@davidradl @twalthr Can we merge this?
Ack, renamed the PR and update the JIRA ticket. The root cause has been fixed with https://github.com/apache/flink/commit/0053b41d5d07252fef468e43c33f30b7fd1e27da.
@flinkbot run azure