Convert Daily Dataframe with Multi Index to quarterly

问题

I would like to convert my daily dataframe of stock data to a quarterly one. However, using resample did not work, because I have a multi index, so I would like my final quarterly dataframe to still contain the individual stocks (resample just summarizes all of them):

import pandas as pd
dict1 = [
        {'ticker':'jpm','date': '2016-11-27','returns': 0.2},
        {'ticker':'jpm','date': '2016-11-28','returns': 0.2},
{'ticker':'ge','date': '2016-11-27','returns': 0.2},
{'ticker':'ge','date': '2016-11-28','returns': 0.2},
{'ticker':'amzn','date': '2016-11-27','returns': 0.2},
{'ticker':'amzn','date': '2016-11-28','returns': 0.2},
]
df1= pd.DataFrame(dict1)
df1['date']      = pd.to_datetime(df1['date'])
df1=df1.set_index(['date','ticker'], drop=True)

My final result is supposed to be:

Q42016 JPM  0.2
Q42016 GE   0.2
Q42016 AMZ  0.2

When I used resample, I get:

Q42016  0.2

Also, I am stuck with Pandas 0.18 (long story). Any help is appreciated.

回答1:

First idea is create DatetimeIndex by convert ticker to column, then use groupby with resample:

df1 = df1.reset_index('ticker').groupby('ticker').resample('Q').mean()
print (df1)
                   returns
ticker date               
amzn   2016-12-31      0.2
ge     2016-12-31      0.2
jpm    2016-12-31      0.2

Another solution with Grouper:

df1 = df1.groupby([pd.Grouper(freq='Q', level='date'), 'ticker']).mean()
print (df1)
                   returns
date       ticker         
2016-12-31 amzn        0.2
           ge          0.2
           jpm         0.2

For quarter periods use assign with to_period and then aggregate by groupby:

df1 = (df1.reset_index()
          .assign(date = lambda x: x['date'].dt.to_period('Q'))
          .groupby(['date','ticker'])
          .mean())
print (df1)
               returns
date   ticker         
2016Q4 amzn        0.2
       ge          0.2
       jpm         0.2

来源：https://stackoverflow.com/questions/58908902/convert-daily-dataframe-with-multi-index-to-quarterly

标签

python

pandas

frequency