pandas - 一次设置多列给出错误 "Not in index error!"

Question

import pandas as pd
df = pd.DataFrame(
    [
        [5, 2],
        [3, 5],
        [5, 5],
        [8, 9],
        [90, 55]
    ],
    columns = ['max_speed', 'shield']
)

df.loc[(df.max_speed > df.shield), ['stat', 'delta']] \
    = 'overspeed', df['max_speed'] - df['shield']

我如上所述使用 .loc 设置多列，在某些情况下我得到 Not in index error!。我在上面做错了什么吗？

Answer 1

在用重复标量overspeed减去之后，创建大小相同的元组列表，如True的数量过滤系列:

m = (df.max_speed > df.shield)
s = df['max_speed'] - df['shield']

df.loc[m, ['stat', 'delta']] = list(zip(['overspeed'] * m.sum(), s[m]))
print(df)
   max_speed  shield       stat  delta
0          5       2  overspeed    3.0
1          3       5        NaN    NaN
2          5       5        NaN    NaN
3          8       9        NaN    NaN
4         90      55  overspeed   35.0

帮助程序 DataFrame 的另一个想法:

df.loc[m, ['stat', 'delta']] = pd.DataFrame({'stat':'overspeed', 'delta':s})[m]

详细信息:

print(list(zip(['overspeed'] * m.sum(), s[m])))
[('overspeed', 3), ('overspeed', 35)]

print (pd.DataFrame({'stat':'overspeed', 'delta':s})[m])
        stat  delta
0  overspeed      3
4  overspeed     35

最简单的就是分别赋值:

df.loc[m, 'stat'] = 'overspeed'
df.loc[m, 'delta'] = df['max_speed'] - df['shield']

print(df)
   max_speed  shield       stat  delta
0          5       2  overspeed    3.0
1          3       5        NaN    NaN
2          5       5        NaN    NaN
3          8       9        NaN    NaN
4         90      55  overspeed   35.0