问题
time A1 A1 A2 A2 A2 A3 A3
2017-01 a1 a2 b1 b2 c .....
2017-02 a3 a4 b3 b4 c
2017-03 a5 a6 b5 b6 c
....
There is a dataframe as shown above. How to get mean value of the columns which have the same name( as shown below)?
time A1 A2 A3
2017-01 (a1+a2)/2 (b1+b2+c)/3 c
2017-02 .....
2017-03
回答1:
Use groupby with level=0 and axis=1.
df.groupby(level=0, axis=1).mean()
np.random.seed(0)
df = pd.DataFrame(np.random.choice(10, (3, 5)), columns=list('AAABB'))
df
A A A B B
0 5 0 3 3 7
1 9 3 5 2 4
2 7 6 8 8 1
df.groupby(level=0, axis=1).mean()
A B
0 2.666667 5.0
1 5.666667 3.0
2 7.000000 4.5
来源：https://stackoverflow.com/questions/53940035/grouping-on-identical-column-names-in-pandas