Github user mgaido91 commented on the issue:
https://github.com/apache/spark/pull/20023
@cloud-fan thanks for your feedback. Honestly I don't have a string opinion
on how we should behave in that case. That's why I wanted some feedbacks from
the community on that point, while this PR covers only the other case, ie.
rounding in case of precision loss in accordance to Hive and SQLServer's
behavior.
Anyway, IIUC now this PR is blocked by #18853 to introduce
`spark.sql.typeCoercion.mode` and use that property to decide whether to keep
the previous behavior or the one proposed here.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
