pandas mean（）赋值空格

Question

我有df3：

                                  1     2    3
A    Battery Voltage, V (Avg)     12    13   14
A    Max Watt, W                  4     5    10
A    Battery Temperature, C       20    21   22

我正在尝试添加一个新列：

M = ['Battery Voltage, V (Avg)', 'Battery Temperature, C']

df3['Total/Max/Min'] = np.where(mask3, df3.mean(axis=1), df3.sum(axis=1))

但我对“平均值”却一片空白

                                  1     2    3     Total/Max/Min
A    Battery Voltage, V (Avg)     12    13   14                   <--
A    Max Watt, W                  4     5    10        19
A    Battery Temperature, C       20    21   22                   <--

编辑： 这两行的type是string，即使值都是数字。我似乎无法转换为float（）

编辑2： 实际测向：

                                           1           2     3        4
CU0111-023349-J Output Energy, (Wh/h)      3.825    1.46    1.57    1.997
CU0111-023349-J Lights (Wh)                 1.53    1.46    1.57    1.997
CU0111-023349-J Lights+Media (Wh)           2.295   0         0       0
CU0111-023349-J Total Usage (h)             1       1         1       1
CU0111-023349-J Lights (h)                  0.4     1         1       1
CU0111-023349-J Light+Media (h)             0.6     0         0       0
CU0111-023349-J Battery Voltage, V (Max)    12.525  12.51   12.502  12.479
CU0111-023349-J Battery Voltage, V (Min)    12.44   12.467  12.421  12.409
CU0111-023349-J Battery Voltage, V (Avg)    12.463  12.494  12.484  12.454
CU0111-023349-J Max Watt, W                 5.213   2.784   3.723   3.615
CU0111-023349-J Battery Temperature, C      19.555  18.977  18.324  17.753
CU0111-023349-J Charge Energy, Wh/h         0.016   0.006   0.009   0.01
CU0111-023349-J Time of Full Charge         11:21   

实际代码：

L = ['Battery Voltage, V (Max)','Max Watt, W']
M = ['Battery Voltage, V (Avg)','Battery Temperature, C']
mask1 = df3.index.get_level_values(1).isin(L)
mask3 = df3.index.get_level_values(1).isin(M)
mask2 = df3.index.get_level_values(1) == 'Battery Voltage, V (Min)'
df3['Total/Max/Min'] = np.where(mask1, df3.max(axis=1),np.where(mask2, df3.min(axis=1),np.where(mask3, df3.median(axis=1), df3.sum(axis=1))))
result:

Answer 1

更新：将所有列转换为数字类型：

df3 = df3.apply(pd.to_numeric, errors='coerce')

---

我无法复制你的问题：

In [182]: df3
Out[182]:
                             1   2   3
A Battery Voltage, V (Avg)  12  13  14
  Max Watt, W                4   5  10
  Battery Temperature, C    20  21  22

In [183]: mask3
Out[183]: [True, True, False]

In [184]: df3['Total/Max/Min'] = np.where(mask3, df3.mean(axis=1), df3.sum(axis=1))

In [185]: df3
Out[185]:
                             1   2   3  Total/Max/Min
A Battery Voltage, V (Avg)  12  13  14      13.000000
  Max Watt, W                4   5  10       6.333333
  Battery Temperature, C    20  21  22      63.000000

你能提供一个可复制的数据集吗？你知道吗