Pandas 将每小时 OHLC 重新采样为每日 OHLC

Question

I have a dataframe of hourly OHLC as follows (please ignore the values of OHLC, I typed them in for better illustration),

hr_df =
                        Close      High       Low      Open
2017-09-04 05:00:00  0.715035  0.715035  0.715035  0.715035
2017-09-04 06:00:00  0.715035  0.715045  0.715015  0.715035
2017-09-04 07:00:00  0.715040  0.715050  0.714035  0.715035
:
:
2017-09-05 05:00:00  0.715045  0.715105  0.714985  0.715035
2017-09-05 06:00:00  0.715040  0.716045  0.714605  0.715035
2017-09-05 07:00:00  0.715040  0.717045  0.713225  0.715035
:
:
2017-09-06 05:00:00  0.715040  0.714045  0.713355  0.715035

I want to resample it into daily OHLC, example,

day_df =
               Close      High       Low      Open
2017-09-04  0.715035  0.715035  0.715035  0.715035
2017-09-05  0.715035  0.715045  0.715015  0.715035
2017-09-06  0.715040  0.715050  0.714035  0.715035
2017-09-07  0.715045  0.715105  0.714985  0.715035
2017-09-08  0.715040  0.716045  0.714605  0.715035
2017-09-09  0.715040  0.714045  0.713355  0.715035
2017-09-10  0.715040  0.717045  0.713225  0.715035

I tried using pandas resample method, day_df = hr_df.resample('D').pad() or day_df = hr_df.resample('D').ohlc() but it is not working. I know I am probably not using the proper method. I will really appreciate it if someone can guide me to an alternative solution or the proper method to use.

Answer 1

I think you need downsample by Resampler.agg by dictionary with keys for column names and values for functions:

day_df = (hr_df.resample('D')
               .agg({'Open': 'first', 'High': 'max', 'Low': 'min', 'Close': 'last'}))
print (day_df)
                Open      High    Close       Low
2017-09-04  0.715035  0.715050  0.71504  0.714035
2017-09-05  0.715035  0.717045  0.71504  0.713225
2017-09-06  0.715035  0.714045  0.71504  0.713355

Answer 2

Try to use pd.Grouper . If pair is a forex pair for example you can use it as index together with the newly grouped date:

hr_df.groupby([pd.Grouper(key='date',freq='D'), 'pair']).agg(
{'Open': 'first', 'High': 'max', 'Low': 'min', 'Close': 'last'}

)

Answer 3

That works for me.

quotes.set_index('end', inplace=True)
quotes.index = pd.to_datetime(quotes.index)
ohlc_dict = {
    'open': 'first',
    'high': 'max',
    'low': 'min',
    'close': 'last'
}
quotes.resample("1D", closed='left', label='left').apply(ohlc_dict).dropna()