Github user hvanhovell commented on the issue:
https://github.com/apache/spark/pull/16818
@uncleGen I think we should limit this to allowing long values for range
frames only; row frames should not get larger than `1 << 31 + 1`. The reason
for this is that we also need to be able to buffer that many rows and that this
currently both not practical (I have yet too see someone hitting this limit),
and that `WindowExec` assumes that the buffers are integer bound (see
`RowBuffer.size` for instance). Also testing this will be a total PITA.
Just make sure we can construct a range frame that respects longs, and
throw an error for row frames.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
