Github user rdblue commented on the issue:
https://github.com/apache/spark/pull/21946
Isn't this unnecessary after the API redesign?
For the redesign, the `DataSourceV2` or a `ReadSupportProvider` will supply
a `create` method (or `anonymousTable`) to return a `Table` that implements
`ReadSupport`. `ReadSupport` should not accept user schemas because the schema
should be accessible from the `Table` itself. That way, we can use the same
table-based relation (see
https://github.com/apache/spark/pull/21877/files#diff-35ba4ffb5ccb9b18b43226f1d5effa23R82).
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
