[
https://issues.apache.org/jira/browse/ARROW-4324?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16752962#comment-16752962
]
Uwe L. Korn commented on ARROW-4324:
------------------------------------
[~wesmckinn] [~pitrou] We regularly get issues like this. We should decide on a
policy what to do with mixed type columns and adjust the error message
accordingly.
I would suggest that we error in the case where we get mixed type columns as
Arrow is focused on a much lower level than pandas and should rather support
type safety than "there is always some output".
> [Python] Array dtype inference incorrect when created from list of mixed
> numpy scalars
> --------------------------------------------------------------------------------------
>
> Key: ARROW-4324
> URL: https://issues.apache.org/jira/browse/ARROW-4324
> Project: Apache Arrow
> Issue Type: Bug
> Affects Versions: 0.11.1
> Reporter: Keith Kraus
> Priority: Minor
>
> Minimal reproducer:
> {code:python}
> import pyarrow as pa
> import numpy as np
> test_list = [np.dtype('int32').type(10), np.dtype('float32').type(0.5)]
> test_array = pa.array(test_list)
> # Expected
> # test_array
> # <pyarrow.lib.DoubleArray object at 0x7f009963bf48>
> # [
> # 10,
> # 0.5
> # ]
> # Got
> # test_array
> # <pyarrow.lib.Int32Array object at 0x7f009963bf48>
> # [
> # 10,
> # 0
> # ]
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
