[
https://issues.apache.org/jira/browse/ARROW-378?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15833575#comment-15833575
]
Jim Ahn commented on ARROW-378:
-------------------------------
No worries. I'm happy to do so, but please consider taking the current pull
request as this 'implicit conversion' to UTC by numpy is documented and
explicitly tested. It may help others in the future avoid being fooled in a
similar manner. :)
> Python: Respect timezone on conversion of Pandas datetime columns
> -----------------------------------------------------------------
>
> Key: ARROW-378
> URL: https://issues.apache.org/jira/browse/ARROW-378
> Project: Apache Arrow
> Issue Type: Bug
> Components: Python
> Reporter: Uwe L. Korn
> Assignee: Jim Ahn
> Labels: newbie
>
> Currently we convert columns irrespectively of their timezone directly to
> timestamps in Arrow. While in Pandas you are able to specify timezones, in
> Arrow we have the assumptions that a Timestamp is always the time of the
> specified unit since 00:00:00.000 on 1 January 1970, UTC. Thus we have to
> check if {{df['column‘].tz}} is set and convert if necessary to UTC before
> converting to an Arrow column.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
