[
https://issues.apache.org/jira/browse/ARROW-3586?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16711424#comment-16711424
]
Francois Saint-Jacques commented on ARROW-3586:
-----------------------------------------------
Segfault was caused by a null deref in
[https://github.com/apache/arrow/blob/apache-arrow-0.11.1/cpp/src/arrow/compute/kernels/util-internal.h#L35]
. This is fixed in master/0.12.0 in
[https://github.com/apache/arrow/blob/master/cpp/src/arrow/array.h#L174] .
> [Python] Segmentation fault when converting empty table to pandas with
> categoricals
> -----------------------------------------------------------------------------------
>
> Key: ARROW-3586
> URL: https://issues.apache.org/jira/browse/ARROW-3586
> Project: Apache Arrow
> Issue Type: Bug
> Affects Versions: 0.10.0, 0.11.0
> Environment: - Ubuntu 16.04, Python 2.7.12, pyarrow 0.11.0, pandas
> 0.23.4
> - Debian9, Python 2.7.13, pyarrow 0.10.0, pandas 0.23.4
> Reporter: Andreas
> Priority: Major
> Fix For: 0.12.0
>
>
> {code:java}
> import pyarrow as pa
> table = pa.Table.from_arrays(arrays=[pa.array([], type=pa.int32())],
> names=['col'])
> table.to_pandas(categories=['col']){code}
> This produces a segmentation fault for certain types (e.g, int\{32,64}) while
> it works for others (e.g. string, binary).
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
