A 1.2  3.6
B 2.6  3.8 
A 1.5  2.1
B 2.0  2.9
C 2.4  3.5
H 4.1  2.9
C 5.9  9.3
A 2.9  8.3
B 8.7  5.6
H 2.9  7.3
H 3.6  1.9
B 2.3  1.0
A 7.1  2.3

import pandas as pd

d = {'col1': ['B','A','B','C','H','C','A','B','H','H','B','B'],'col2': [2.6,1.5,2.0,2.4,4.1,5.9,2.9,8.7,2.9,3.6,2.3,2.3], 'col3': [3.8,2.1,2.9,3.5,2.9,9.3,8.3,5.6,7.3,1.9,1,1]}
df = pd.DataFrame(data=d)
df_avg=df.groupby(['col1']).mean() # this line works
df_pct=df.groupby(['col1']).pct_change.sum()

df_pct=df.groupby(['col1']).pct_change().sum()

The above line of code does only giving the pct_change sum by column (but I want to by column & by group).

The below output is of: df_avg=df.groupby(['col1']).mean()

col1	col2	col3
A	    2.20	5.20
B	    3.57	2.86
C	    4.15	6.40
H	    3.50	4.03

Similarly I want pct_change().sum() by group

df_pct=df.groupby(['col1']).pct_change().sum()