Pandas dataframe dates off by one day

Question

I noticed that if I have pandas datetime index inside a pandas dataframe, the value of the dates changes when I call them by using .value

Example.

import pandas as pd

date_try=pd.date_range(start='2017-01-01',end='2017-05-01',freq='MS')
date_df=pd.DataFrame(date_try)

print(date_df)

enter image description here

Now notice how the date values change when I call

print(date_df.values)

enter image description here

What version of pandas do you have? I cannot reproduce this on 0.20 — cs95, Jul 17 '17 at 17:34
My version is 0.15.2 and I am restricted to using this version. — pandas_get_good, Jul 17 '17 at 17:43
@MaxU Well, its giving me trouble since its giving me a wrong date. So I am hoping there is a way to get around this — pandas_get_good, Jul 17 '17 at 17:44
What you see is the difference between GMT and your local time (time zone -8 in winter, -7 in summer). When it is midnight of 2017-01-01 in Greenwich, it is still 4PM of 2016-12-31 locally. — DYZ, Jul 17 '17 at 17:44

score 2 · Accepted Answer · answered Jul 17 '17 at 17:54

@DYZ is right, I run into this all the time setting up cronjobs on my aws box, where the time it's run doesn't correspond to what the source-file says is the time because it was produced in a different timezone You can try to re-index it in local time.

import pytz
eastern = pytz.timezone('US/Eastern')
df.index = index.tz_localize(pytz.utc).tz_convert(eastern)

Pandas dataframe dates off by one day

1 Answers1