python - python pandas groupby：key1的key2值的百分比

Question

我有一个数据框，并按两个键对其进行分组df.groupby(['key1',key2'])。对于每个 key2 条目，如何显示其 key1 值的百分比？

Answer 1

Here's an alternative method using one groupby statement.

Group by k1, select column k2 and apply a lambda function. The lambda gets frequency counts for each level of k2 within k1 and then we divide by the count of k1:

In [1]: df.groupby('k1')['k2'].apply(lambda x: pd.value_counts(x)/x.count().astype(float))

Out[1]:
k1
a   x    0.500000
    y    0.500000
b   y    0.666667
    x    0.333333

Performance:

HYRY's method:

100 loops, best of 3: 3.07 ms per loop

My method:

1000 loops, best of 3: 1.98 ms per loop

Answer 2

为“k1”和（“k1”，“k2”）调用 groupby 两次，然后执行div：

import pandas as pd
k1 = ["a", "a", "a", "a", "b", "b", "b"]
k2 = ["x", "x", "y", "y", "x", "y", "y"]
df = pd.DataFrame({"k1":k1, "k2":k2})

df.groupby(["k1", "k2"]).k2.count().div(
    df.groupby("k1").k1.count().astype(float), level=0)

输出：

k1  k2
a   x     0.500000
    y     0.500000
b   x     0.333333
    y     0.666667