Github user heary-cao commented on the issue:
https://github.com/apache/spark/pull/18069
@hvanhovell
thank you for review it.
This is an extreme case. must be with `Longtype` type and `Decimal` type
mixing test.
However, when test case:
spark-sql> select (12345678901234
Github user hvanhovell commented on the issue:
https://github.com/apache/spark/pull/18069
I don't think this makes any sense:
- The parser is not the place where you should make this change. This
should be done in `TypeCoercion`. Now we create different semantics between SQL