如何使用Python计算列指定值占比并按国家分组聚合统计

我们可以用pandas库完成该统计需求，完整实现代码如下：

import pandas as pd

# 构造示例数据表，如果你已经有对应DataFrame可跳过该步骤
data = [
    ["US", "a1", "upload"],
    ["US", "a1", "delete"],
    ["US", "a1", "delete"],
    ["US", "a2", "upload"],
    ["US", "a2", "upload"],
    ["US", "a2", "delete"],
    ["IT", "k2", "upload"],
    ["IT", "k2", "upload"],
    ["IT", "k2", "delete"],
]
df = pd.DataFrame(data, columns=["country", "ID", "type"])

# 第一步：按国家+ID分组，计算每个ID的delete占比
id_level_stats = df.groupby(["country", "ID"])["type"].agg(
    delete_ratio=lambda x: (x == "delete").sum() / len(x)
).reset_index()

# 第二步：按国家分组，统计唯一ID数、所有ID delete占比的平均值
result = id_level_stats.groupby("country").agg(
    ID_cnt=("ID", "nunique"),
    delete_perc=("delete_ratio", lambda x: round(x.mean(), 2))
).reset_index()

# 输出结果
print(result)

运行后输出的result就是你需要的统计结果：

如果不需要强制保留两位小数，把第二步中delete_perc对应的聚合规则lambda x: round(x.mean(), 2)换成'mean'即可。

内容的提问来源于stack exchange，提问作者french_fries