pandas get the mean len of columns that is a list in groupby-CodePudding

I have a dataframe:

df = c    l 
     A  [1,2,3]
     A  [1,2,3,4]
     B  [1]

I want to get the mean length (and std) of the column l, per group of c. So here the output will be: A : 3.5 B : 1

What is the best way to do so?

CodePudding user response：

Use Series.str.len with aggregate mean:

s = df['l'].str.len().groupby(df['c']).mean()
print (s)
c
A    3.5
B    1.0
Name: l, dtype: float64

Or with aggregate multiple functions:

df = df['l'].str.len().groupby(df['c']).agg(['mean','std'])
print (df)
   mean       std
c                
A   3.5  0.707107
B   1.0       NaN