[
https://issues.apache.org/jira/browse/ARROW-1890?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16280148#comment-16280148
]
ASF GitHub Bot commented on ARROW-1890:
---------------------------------------
xhochy commented on a change in pull request #1393: ARROW-1890: [Python] Fix
mask handling for Date32 NumPy conversions
URL: https://github.com/apache/arrow/pull/1393#discussion_r155229720
##########
File path: python/pyarrow/tests/test_convert_pandas.py
##########
@@ -572,6 +572,15 @@ def test_date_infer(self):
expected['date'] = pd.to_datetime(df['date'])
tm.assert_frame_equal(result, expected)
+ def test_date_mask(self):
+ arr = np.array([date(2017, 4, 3), date(2017, 4, 4)],
dtype='datetime64[D]')
+ mask = [True, False]
+ t32 = pa.date32()
+ result = pa.array(arr, mask=np.array(mask))
+ expected = np.array([None, date(2017, 4, 4)], dtype='datetime64[D]')
+ expected = pa.array(expected, from_pandas=True)
Review comment:
I would not expected that we would need `from_pandas=True` here so that None
turns into NaT instead of 1970 but this is a different issue:
https://issues.apache.org/jira/browse/ARROW-1891
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
> [Python] Masking for date32 arrays not working
> ----------------------------------------------
>
> Key: ARROW-1890
> URL: https://issues.apache.org/jira/browse/ARROW-1890
> Project: Apache Arrow
> Issue Type: Bug
> Reporter: Uwe L. Korn
> Assignee: Uwe L. Korn
> Labels: pull-request-available
>
> {code}
> arr = np.array([date(2017, 4, 3), date(2017, 4, 4)], dtype='datetime64[D]')
> mask = [True, False]
> t32 = pa.date32()
> result = pa.array(arr, mask=np.array(mask))
> expected = pa.array(np.array([None, date(2017, 4, 4)], dtype='datetime64[D]'))
> {code}
> The above code produces an array that has no null values.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
