jorisvandenbossche commented on issue #15047:
URL: https://github.com/apache/arrow/issues/15047#issuecomment-1593550760
The fallback mentioned above is about the conversion pyarrow ->
python/pandas. In the past, we required pytz for this, but now in an
environment where pytz is not installed, you can see we use zoneinfo. Using
your example `ts`:
```
>>> ts
<pyarrow.lib.TimestampArray object at 0x7f8c9f762ec0>
[
2038-04-01 10:00:00.000000
]
>>> ts[0].as_py()
datetime.datetime(2038, 4, 1, 4, 0,
tzinfo=zoneinfo.ZoneInfo(key='America/Boise'))
```
which uses a ZoneInfo. While if I then install pytz in that environment, the
result I get is:
```
>>> ts[0].as_py()
datetime.datetime(2038, 4, 1, 3, 0, tzinfo=<DstTzInfo 'America/Boise' MST-1
day, 17:00:00 STD>)
>>> type(ts[0].as_py().tzinfo)
<class 'pytz.tzfile.America/Boise'>
```
Now, the conversion from python -> pyarrow should already support zoneinfo,
regardless of pytz being available (the function that handles the timezone has
several cases depending on the exact tz object). But it might be there is a bug
in there (certainly something to investigate, clearly!), but so that's a
separate issue. Can you open a new issue for that?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
