Python计算年龄标准化率时出现NaN值问题求助

我在求职项目中处理疾病相关数据，前期计算乌干达和美国的分年龄人口权重百分比时结果正常，无NaN值。但后续用COPD表中的死亡率乘以对应权重计算年龄标准化率时，部分行出现NaN值。已排除除零情况，仅新计算的列存在该问题。

相关代码与数据 #

无错误的权重计算代码 #

# Calculation of the weight of the population by age

#Uganda
# Sum of observations from the column "Estimates, Uganda, 2019"
total_observation = transposed_df["Estimates, Uganda, 2019"].sum()

# The weight in percentage for each observation
transposed_df["Weight Percentage Uganda"] = (transposed_df["Estimates, Uganda, 2019"] / total_observation)

##### 
#United States of America

# Sum of observations from the column "Estimates, United States of America, 2019"
total_observation1 = transposed_df["Estimates, United States of America, 2019"].sum()

# The weight in percentage for each observation
transposed_df["Weight Percentage United States of America"] = (transposed_df["Estimates, United States of America, 2019"] / total_observation1)

#print
transposed_df

出现NaN的年龄标准化率计算代码 #

#Age-standardized rate from Uganda 

transposed_df["Age-standardized rate from Uganda"] = (COPD["Death rate, Uganda, 2019"] * transposed_df["Weight Percentage Uganda"])

#Age-standardized rate from United States of America

transposed_df["Age-standardized rate from United States of America"] = (COPD["Death rate, United States, 2019"] * transposed_df["Weight Percentage United States of America"])

#display result
transposed_df

COPD数据表 #

排查情况：仅新计算的年龄标准化率列出现NaN，前期权重计算正常，已排除除零问题。

可能的原因及解决方案 #

1. 索引不匹配（最常见原因） #

如果COPD和transposed_df的行索引未基于相同年龄组，或索引顺序不一致，相乘时会因无法匹配对应行产生NaN。

解决方法：统一两个DataFrame的索引为年龄组，确保对齐：

# 将COPD的索引设置为年龄组
COPD.set_index("Age group (years)", inplace=True)
# 确保transposed_df的索引也是年龄组（若未设置）
transposed_df.set_index("Age group (years)", inplace=True)

# 重新计算年龄标准化率
transposed_df["Age-standardized rate from Uganda"] = COPD["Death rate, Uganda, 2019"] * transposed_df["Weight Percentage Uganda"]
transposed_df["Age-standardized rate from United States of America"] = COPD["Death rate, United States, 2019"] * transposed_df["Weight Percentage United States of America"]

2. COPD表存在隐藏NaN值 #

展示的数据表无NaN，但实际加载的COPD数据可能存在缺失值。先检查：

# 检查COPD表中的NaN值
print(COPD.isna().sum())

若发现NaN，可根据需求填充：

# 用0填充NaN
COPD.fillna(0, inplace=True)
# 或用列均值填充
COPD.fillna(COPD.mean(), inplace=True)

3. 数据类型不兼容 #

若死亡率列为字符串类型，相乘时会自动转为NaN。检查并转换数据类型：

# 查看COPD表的数据类型
print(COPD.dtypes)

# 转换为数值型
COPD["Death rate, United States, 2019"] = pd.to_numeric(COPD["Death rate, United States, 2019"], errors="coerce")
COPD["Death rate, Uganda, 2019"] = pd.to_numeric(COPD["Death rate, Uganda, 2019"], errors="coerce")

4. 手动对齐行顺序 #

若索引设置后仍有问题，可按年龄组排序确保行顺序一致：

# 按年龄组排序
COPD.sort_index(inplace=True)
transposed_df.sort_index(inplace=True)

内容的提问来源于stack exchange，提问作者Giovanna