python - Pandas 按一列的类别在所有表中出现唯一值的频率

Question

有一列有4个类别，我想为每个唯一值显示其他列的值出现的频率部分数据示例

输出

Answer 1

从这里开始:

df = pd.DataFrame(
    {
        "cat1": ["yes", "no", "yes", "no", "yes"],
        "cat2": ["a", "a", "b", "b", "a"],
        "cat3": ["yes", "no", "no", "yes", "no"],
        "quant": [1, 2, 3, 4, 5],
    }
)

示例数据框:

    cat1 cat2 cat3  quant
0   yes  a    yes   1
1   no   a    no    2
2   yes  b    no    3
3   no   b    yes   4
4   yes  a    no    5

你可以这样做:

y = lambda x: x.value_counts(normalize=True).loc["yes"]
n = lambda x: x.value_counts(normalize=True).loc["no"]
df.groupby(["cat2"]).agg(
    {
        "cat1": [("yes", y), ("no", n)],
        "cat3": [("yes", y), ("no", n)],
        "quant": ["min", "max", "mean"],
    }
)

结果:

      cat1                  cat3                   quant
      yes           no      yes         no         min  max mean
cat2                            
a     0.666667  0.333333    0.333333    0.666667    1   5   2.666667
b     0.500000  0.500000    0.500000    0.500000    3   4   3.500000

这是一个稍微更健壮的版本:

from functools import partial

def agg_func(s: pd.Series, name: str):
    try:
        return s.value_counts(normalize=True).loc[name]
    except KeyError:
        return 0


yes_no_agg = [
    ("yes", partial(agg_func, name="yes")),
    ("no", partial(agg_func, name="no")),
]

df.groupby(["cat2"]).agg(
    {
        "cat1": yes_no_agg,
        "cat3": yes_no_agg,
        "quant": ["min", "max", "mean"],
    }
)