How to rename timestamp column names to string/object in multiindex dataframe using python

Question

DataFrame :df

None  | A  B   volumeshare    volumeshare
X     |         2020-10-1      2020-11-1
---------------------------------------
0     | e1 f1      12            65    
1     | e1 f2      23            20     
2     | e1 f3      0             91    
3     | e2 f1      76            3     
4     | e2 f2      89            33    

I wish to achieve the below things -

1. Only one top row with X, A, B, 2020-10-1, 2020-11-1 as the column names in that order.
2. column 3 and 4 have timestamp names (2020-10-1 & 2020-11-1), need to replace it into string as Oct-20 and Nov-20 respectively.

Expected output

X     | A   B      Oct-20       Nov-20
---------------------------------------
0     | e1 f1      12            65    
1     | e1 f2      23            20     
2     | e1 f3      0             91    
3     | e2 f1      76            3     
4     | e2 f2      89            33    

Answer 1

Use custom lambda function in list comprehension:

def f(a, b):
    d = pd.to_datetime(b, errors='coerce')
    return d.strftime('%b-%y') if pd.notna(d) else a
   
df.columns = [f(a, b) for a, b in df.columns]
print (df)
    A   B  Oct-20  Nov-20
0  e1  f1      12      65
1  e1  f2      23      20
2  e1  f3       0      91
3  e2  f1      76       3
4  e2  f2      89      33