CODEWITHSUNDEEP
pythonpandasdayofweek
anonaka
anonaka

Reputation: 95

How can I get day_name() after groupby dayofweek?

I have a DatetimeIndex indexed Series s1 like

s1

time
2021-08-24 09:24:16+09:00    11933142
2021-08-24 10:00:03+09:00    11785209
2021-08-24 11:00:03+09:00    14462866
2021-08-24 19:00:04+09:00    11419204
2021-08-24 20:00:03+09:00    11757634
Name: x, dtype: int64

and want mean() for each day of the week. Then I want to get day_name() of the index to draw a graph label but I cannot since DatetimeIndex now becomes int.

g1 = s1.groupby(s1.index.dayofweek).mean()
----> 3 g1.index.day_name()

AttributeError: 'Int64Index' object has no attribute 'day_name'


What is the smartest workaround?

Upvotes: 1

Views: 149

Answers (2)

Corralien
Corralien

Reputation: 120401

Group by day_name:

>>> s1.groupby(s1.index.day_name()).mean()

Tuesday    12271611.0
Name: time, dtype: float64

Graph example:

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import calendar

days = list(calendar.day_name)

s2 = pd.Series(np.random.randint(1000, 10000, 365),
               index=pd.date_range('2021-01-01', '2021-12-31'))

s2.groupby(s2.index.day_name()).mean().reindex(days).plot(kind='bar', rot=45)

plt.show()

Graph example

Upvotes: 1

U13-Forward
U13-Forward

Reputation: 71570

Try this:

>>> s1.groupby(s1.index.dayofweek).transform('mean').index.day_name().drop_duplicates()
Index(['Tuesday'], dtype='object')
>>>

Upvotes: 0

Related Questions