通过 object 在一个 group 的 pandas agg 方法中传递一个 function

Question

This is in google colab

import pandas as pd
df = pd.DataFrame({'A': ['state1', 'state2', 'state1', 'state1', 'state2'], 
                   'B': [1, 2, 3, 4, 5], 
                   'C': [10, 11, 12, 13, 14]})
print(df)

      A     B    C
0   state1  1   10
1   state2  2   11
2   state1  3   12
3   state1  4   13
4   state2  5   14

This is the DataFrame. If we use the built-in 'max' function with the agg function it is working fine.

df.groupby('A', as_index=False)['C'].agg('max')

      A     C
0   state1  13
1   state2  14

But if we use a lambda function is raising an error

df.groupby('A', as_index=False)['C'].agg(lambda x: x.max())

AttributeError: 'Series' object has no attribute 'columns'

Why is this raising a error? What is the difference between passing a lambda function and a built-in function in the agg method?

Even for me it is not working in jupyter notebook also

Answer 1

This works for me:

df.groupby('A', as_index=False)[['C']].agg(lambda x: x.max())

Answer 2

Previously when I run the question for the first time the pandas version was 1.0.5.

But after then i updated the version to 1.1.0 and ran the same thing and it works fine.

!pip install pandas==1.1.0
import pandas as pd

df = pd.DataFrame({'A': ['state1', 'state2', 'state1', 'state1', 'state2'], 
                   'B': [1, 2, 3, 4, 5], 
                   'C': [10, 11, 12, 13, 14]})

print(df.groupby('A', as_index=False)['C'].agg(lambda x: x.max()))

       A     C
0   state1  13
1   state2  14