熊猫数据框通过一些值填充答案

【问题标题】：Pandas dataframe fillna by some value熊猫数据框通过一些值填充
【发布时间】：2019-01-05 07:58:15
【问题描述】：

我有这个数据

import numpy as np
import pandas as pd
group = {'gender': ['male', 'female', 'female', 'male', 'female', 'male', 'male'],
        'height': [175, 168, np.nan, 170, 167, np.nan, 190],
        }
labels = ['a', 'b', 'c', 'd', 'e', 'f', 'g']
df = pd.DataFrame(group, index=labels)
df2 = df.groupby('gender')['height'].mean()

我想用 df2 的平均值填充 nan

【问题讨论】：

标签： python pandas numpy dataframe pandas-groupby

【解决方案1】：

代码

import pandas as pd
import numpy as np

group = {'gender': ['male', 'female', 'female', 'male', 'female', 'male', 'male'],
        'height': [175, 168, np.nan, 170, 167, np.nan, 190],
        }
labels = ['a', 'b', 'c', 'd', 'e', 'f', 'g']
df = pd.DataFrame(group, index=labels)
df2 = df.groupby('gender')['height'].mean()
df['height'].fillna(df['gender'].map(df2), inplace=True)
# print(df2)
print(df)

输出

   gender      height
a    male  175.000000
b  female  168.000000
c  female  167.500000
d    male  170.000000
e  female  167.000000
f    male  178.333333
g    male  190.000000

【讨论】：

【解决方案2】：

您可以使用groupby + transform 和mean。然后fillna 与结果系列。

means = df.groupby('gender')['height'].transform('mean')
df['height'] = df['height'].fillna(means)

print(df)

   gender      height
a    male  175.000000
b  female  168.000000
c  female  167.500000
d    male  170.000000
e  female  167.000000
f    male  178.333333
g    male  190.000000

【讨论】：