Pandas groupby 和自定义 agg lambda function

Question

I have a data frame and I want to aggregate a custom aggregation function.

Right now I have it like a predefined function, but I want to call it as a lambda function. Notice that the predefined function has a parameter that can be change.

from sklearn.datasets import load_boston
import pandas as pd
import numpy as np
bunch = load_boston()

y = bunch.target
X = pd.DataFrame(bunch.data, columns=bunch.feature_names)

def percentile_func(y,PERCENTILE=50):
    return np.percentile(y,PERCENTILE)

X.groupby('CHAS')['CRIM'].agg([percentile_func,'sum', 'count'])

Answer 1

Try this:

X.groupby('CHAS')['CRIM'].agg([lambda x: np.percentile(x, 50),'sum', 'count'])

Answer 2

You can use functools.partial here.

from functools import partial

f = partial(percentile_func,PERCENTILE=50) # you can change PERCENTILE value accordingly.

X.groupby('CHAS')['CRIM'].agg([f,'sum', 'count'])