[
https://issues.apache.org/jira/browse/FLINK-7137?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16093188#comment-16093188
]
ASF GitHub Bot commented on FLINK-7137:
---------------------------------------
Github user twalthr commented on the issue:
https://github.com/apache/flink/pull/4314
I tried to improve the stability regarding nullable types. This should fix
FLINK-7177 and FLINK-7137. It ensures that all types have the correct
nullability. By default, types are always nullable. If they are not nullable,
than we know it for sure.
https://github.com/apache/flink/compare/master...twalthr:FLINK-7137
What do you think? @walterddr @fhueske
> Flink table API defaults top level fields as nullable and all nested fields
> within CompositeType as non-nullable
> ----------------------------------------------------------------------------------------------------------------
>
> Key: FLINK-7137
> URL: https://issues.apache.org/jira/browse/FLINK-7137
> Project: Flink
> Issue Type: Bug
> Components: Table API & SQL
> Affects Versions: 1.3.1
> Reporter: Rong Rong
> Assignee: Rong Rong
>
> Right now FlinkTypeFactory does conversion between Flink TypeInformation to
> Calcite RelDataType by assuming the following:
> All top level fields will be set to nullable and all nested fields within
> CompositeRelDataType and GenericRelDataType will be set to Calcite default
> (which is non-nullable).
> This triggers Calcite SQL optimization engine drop all `IS NOT NULL` clause
> on nested fields, and would not be able to optimize when top level fields
> were actually non-nullable.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
