groupby(…).agg(…)

How to use built-in or custom functions in groupby operations and set the name of the resulting column.

import pandas as pd
import numpy as np

# use iris dataset from seaborn package
df = pd.read_csv(
    "https://raw.githubusercontent.com/mwaskom/seaborn-data/master/iris.csv"
)

df_grouped = (
    df.groupby("species")
    .agg(
        pl_min=("petal_length", "min"),
        pl_max=("petal_length", lambda x: max(x)),
        pl_perc95=("petal_length", lambda x: np.percentile(x, 95)),
    )
    .reset_index()
)

df_grouped

	species	pl_min	pl_max	pl_perc95
0	setosa	1.0	1.9	1.700
1	versicolor	3.0	5.1	4.900
2	virginica	4.5	6.9	6.655

Aggregate all rows

import pandas as pd
import numpy as np

# use iris dataset from seaborn package
df = pd.read_csv(
    "https://raw.githubusercontent.com/mwaskom/seaborn-data/master/iris.csv"
)

df_grouped = (
    df.groupby(lambda x: True)  # from https://stackoverflow.com/a/46583472
    .agg(
        pl_min=("petal_length", "min"),
        pl_max=("petal_length", lambda x: max(x)),
        pl_perc95=("petal_length", lambda x: np.percentile(x, 95)),
    )
    .reset_index()
)

df_grouped

	index	pl_min	pl_max	pl_perc95
0	True	1.0	6.9	6.1