Python 基于直方图箱的样本数据

您可以使用cut来形成bin，然后计算Cable_ind和bin的crosstab，得到每个bin的min.最后，使用sample调整数据帧，并使用cumcount获得最多可能值的数量:

data = {
    'size': [10, 10, 10, 10, 10, 10, 10, 20, 20, 20, 20, 20, 30, 30, 40, 40, 50, 50, 60, 60, 60],
    'cable_ind': [0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 1, 1, 1, 0, 1, 1, 1, 0, 1, 1, 0]
}

df = pd.DataFrame(data)

# compute bins
bin_s = pd.cut(df['size'], bins=[10, 20, 30, 40, 50, 60, 70], right=False)

# compute crosstab, determine the max number of items to keep per bin
keep = pd.crosstab(df['cable_ind'], bin_s).min()

# shuffle (optional), get up to the max number of common items per bin
out = df[df.sample(frac=1) # only if you want a random sampling
           .groupby(['cable_ind', bin_s])
           .cumcount().lt(bin_s.map(keep))]

输出示例(为清楚起见，以bin_s为列):

    size  cable_ind     bin_s
0     10          0  [10, 20)
6     10          1  [10, 20)
7     20          1  [20, 30)
8     20          0  [20, 30)
12    30          1  [30, 40)
13    30          0  [30, 40)
16    50          1  [50, 60)
17    50          0  [50, 60)
19    60          1  [60, 70)
20    60          0  [60, 70)

输出图表:

sns.histplot(data=out, x='size', hue='cable_ind',
             bins=[10, 20, 30, 40, 50, 60, 70],
             multiple='stack', edgecolor='black')