Mingcong Han created SPARK-27572:
------------------------------------
Summary: IN predicate should be nullable
Key: SPARK-27572
URL: https://issues.apache.org/jira/browse/SPARK-27572
Project: Spark
Issue Type: Bug
Components: SQL
Affects Versions: 3.0.0
Reporter: Mingcong Han
Currently, we are not dealing with the nullability of IN predicate correctly.
1. IN(In, InSet, InSubquery) with nulls in the left values should return FALSE
if the right side is an empty set. Example: `null in (select a from t where
false)` should return false.
2. InSubquery is nullable according to the SQL standard. But we are using a
non-nullable ExistenceJoin for the expression `(a in (select b from t)) is
null`.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
