You are viewing a plain text version of this content. The canonical link for it is here.
Posted to jira@arrow.apache.org by "Alessandro Molina (Jira)" <ji...@apache.org> on 2021/10/20 13:41:00 UTC
[jira] [Commented] (ARROW-14388) [Python] Add unittests for
converter arrays with pandas masks
[ https://issues.apache.org/jira/browse/ARROW-14388?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17431246#comment-17431246 ]
Alessandro Molina commented on ARROW-14388:
-------------------------------------------
[~jorisvandenbossche] [~kszucs] I was able to reproduce the segfault using the provided test. I confirmed the test reproduces the issue by reverting the changes in [https://github.com/apache/arrow/pull/11465] and triggering the segfault.
I also added an additional check, that I verified does prevent the segfault replacing it with a proper {{Invalid("Invalid mask type")}} error to catch future regressions.
I couldn't find a way to trigger that error with current codebase on master, so it ends up being uncovered by a test. That's because with current codebase anything that is not a numpy.array gets converted to it so there is no way to end up into that situation normally. Ideally it's a kind of issue I would simulate by monkeypatching, but given that everything runs within Cython I can't monkeypatch {{get_values}} anyway the check is there and should prevent us from reintroducing the same issue in the future.
> [Python] Add unittests for converter arrays with pandas masks
> -------------------------------------------------------------
>
> Key: ARROW-14388
> URL: https://issues.apache.org/jira/browse/ARROW-14388
> Project: Apache Arrow
> Issue Type: Improvement
> Components: Python
> Reporter: Krisztian Szucs
> Assignee: Alessandro Molina
> Priority: Major
> Fix For: 7.0.0
>
>
> Cover the changes in https://github.com/apache/arrow/pull/11465
> cc [~amol-]
--
This message was sent by Atlassian Jira
(v8.3.4#803005)