Details
-
Bug
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
None
Description
Somewhat related to ARROW-6386 on the interpretation of nulls, we currently have the following behaviour:
In [28]: pa.NULL == pa.NULL Out[28]: True
Which I think is certainly unexpected for a null / missing value. I still need to check what the array-level compare kernel does (NULL or False? ideally NULL I think), but we should follow that.
Attachments
Issue Links
- links to