I can't convert 1H OHLC data into monthly data with a specific time by loffset

[yomoko]

#13218
#13933
With regard to the above issues, I have the same problem. I need to convert 1 hour to monthly data, the requirement is that I need to set the beginning time interval to 17:00.

ohlc_dict = {'open':'first', 'high':'max', 'low':'min', 'close': 'last'}
df = df.resample('1M', loffset='17H',closed='left',label='left').agg(ohlc_dict)
I specified the time by loffset='17H' already, it just changed time label to 17:00, but it still took the Open Price from 00:00 instead of 17:00. What can I do to the get open price at 17:00 instead of 00:00? Thanks.

[chris-b1]

please show a reproducible example

[yomoko]

@TomAugspurger Let's make it simpler. If I am resampling from 1 hour to 1 month, is it possible to specify a base with a different unit such as hours?

[TomAugspurger]

Let us know if you can make an example, http://matthewrocklin.com/blog/work/2018/02/28/minimal-bug-reports

[andreivirtosu]

see reproducible example below:

>>> df = pd.DataFrame(np.random.rand(60,2), columns=list('AB'), index=pd.date_range('2010-01-01 09:00:00', periods=60, freq='1min'))
>>> df.head()
                            A         B
2010-01-01 09:00:00  0.182271  0.367963
2010-01-01 09:01:00  0.562164  0.465836
2010-01-01 09:02:00  0.836276  0.905308
2010-01-01 09:03:00  0.397262  0.216895
2010-01-01 09:04:00  0.637719  0.182088
>>> df.resample('15min').agg({'A':'first','B':'last'})
                            A         B
2010-01-01 09:00:00  0.182271  0.759560
2010-01-01 09:15:00  0.187075  0.563832
2010-01-01 09:30:00  0.303773  0.302590
2010-01-01 09:45:00  0.531070  0.493919
>>> df.resample('15min', loffset='5min').agg({'A':'first','B':'last'})
                            A         B
2010-01-01 09:05:00  0.182271  0.759560
2010-01-01 09:20:00  0.187075  0.563832
2010-01-01 09:35:00  0.303773  0.302590
2010-01-01 09:50:00  0.531070  0.493919

Adding the loffset changed the timestamp but the A B columns values were left unchanged.