Github user cloud-fan commented on the issue:
https://github.com/apache/spark/pull/20023
it looks like other databases are very careful about precision lose.
However, following DB2 and throwing exception is pretty bad for big data
applications, but returning null is also bad as it violates SQL standard.
A new proposal: can we increase the max decimal precision to 76 and keep
max scale as 38? Then we can avoid precision lose IIUC.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
