I have a pandas dataframe with timestamps, like that:
df = pd.DataFrame({'a': [pd.Timestamp('2020-01-01 00:00:00')]})
I write it to a feather file with df.to_feather('a.feather')
, and read it back in R with df <- arrow::read_feather('a.feather')
.
When I display it, I see
A tibble: 1 × 1 a
<dttm>
2020-01-01 01:00:00
Where did the 01:00:00
come from? How can I get rid of it?
ok, it seems that R or arrow is paying attention to some kind of global time zone variable. So the above behaviour can be fixed by adding
Sys.setenv(TZ='GMT')
in the beggining.
Inspired by this: https://arrow.apache.org/docs/python/timestamps.html