zhengruifeng commented on PR #36181: URL: https://github.com/apache/spark/pull/36181#issuecomment-1103419993
@xinrong-databricks I think we should keep in line with the pandas on dealing with NAs, and end users do not need to know the internal details about converting NAs to nulls. For this PR, I think we may impl `skipna` in this way: 1, when `skipna=True`, we keep current behavior which converts pandas' NA (None, np.nan) to `null` in SQL; 2, when `skipna=False`, we add an extra conversion which convert `null` to `Double.NaN`; 3, if there are some statistics which can not support `skipna` in above way, we may need to add some dedicated functions in the SQL side. also cc @HyukjinKwon @itholic @ueshin -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
