如何使用一系列值来估算/替换 Pandas DataFrame 中的缺失值？

Question

For example, suppose I have this DataFrame:

Weights = pd.DataFrame({'Weight': [46, np.nan, 67, 62, np.nan, np.nan, 88, np.nan, 55, np.nan]})
Weights

    Weight
0   46.0
1   NaN
2   67.0
3   62.0
4   NaN
5   NaN
6   88.0
7   NaN
8   55.0
9   NaN

And I would like to replace/impute the NaN values with the following sequence of values:

replace = np.random.randint(45,90, size=(5,))
replace
array([85, 79, 68, 72, 52])

Such that the resulting DateFrame looks like:

Weights

    Weight
0   46
1   85
2   67
3   62
4   79
5   68
6   88
7   72
8   55
9   52

What code do I need? Could this be done using standard python code, only pandas, or only scikit-learn? Thanks in advance.

Answer 1

我刚刚想通了

weights.replace({np.nan:replace})