Convert datetime columns to a different timezone pandas

Question

I have two datetime columns which are naive when I read them into memory but which are in US/Eastern actually. I simply want to convert both of these columns to US/Central.

I found a method which works but it seems like I am doing a workaround. I changed my call_start and call_end columns to be named 'start' and 'end' instead so I don't end up with duplicate column names. I then created a separate datetimeindex for each of these columns and reset the index.

aht.set_index(pd.DatetimeIndex(aht['start']).tz_localize('US/Eastern').tz_convert('US/Central'), inplace = True, drop = True)
aht.index.names = ['call_start']
aht = aht.reset_index()
aht.set_index(pd.DatetimeIndex(aht['end']).tz_localize('US/Eastern').tz_convert('US/Central'), inplace = True, drop = True)
aht.index.names = ['call_end']
aht = aht.reset_index()

I end up getting:

                 call_end                  call_start              start                 end
2016-01-13 06:05:01-06:00   2016-01-13 06:02:00-06:00   01/13/2016 07:02    01/13/2016 07:05
2016-01-13 06:07:00-06:00   2016-01-13 06:03:16-06:00   01/13/2016 07:03    01/13/2016 07:07
2016-01-13 06:09:13-06:00   2016-01-13 06:06:02-06:00   01/13/2016 07:06    01/13/2016 07:09
2016-01-13 06:17:51-06:00   2016-01-13 06:06:20-06:00   01/13/2016 07:06    01/13/2016 07:17

Is this the best method? All other data is in central time so I just want to make sure that this file is too so when I merge files together it makes more sense. I do not care about having the actual timezone stamp there though - is there a way to easily strip it after I created my new columns?

Community · Accepted Answer · 2017-05-23 12:31:35Z

67

You don't need to do the roundtrip to DatetimeIndex, as these methods are avaliable for a Series (column) as well through the dt accessor:

aht['call_start'] = aht['start'].dt.tz_localize('US/Eastern').dt.tz_convert('US/Central')

And the same for end.
To remove the timezone information but keep it in the local time, you do another .dt.tz_localize(None) afterwards (see this question: https://stackoverflow.com/a/34687479/653364)

edited May 23, 2017 at 12:31

CommunityBot

11 silver badge

answered Jan 14, 2016 at 13:01

joris

140k37 gold badges258 silver badges207 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

trench Over a year ago

Well I am glad you answered - I literally tried the DatetimeIndex method because of this comment by you haha. I never messed with DatetimeIndexes before. I'll try the accessor now. "Note: you can't use tz_localize or tz_convert directly on a column, only on an index – joris Sep 3 '14 at 22:41 "

joris Over a year ago

Situation has changed since then :-) The datetime accessor was added in pandas 0.15, exactly to more easily access these DatetimeIndex methods for datetime data in a Series.

Collectives™ on Stack Overflow

Convert datetime columns to a different timezone pandas

1 Answer 1

2 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related