python - 为什么 pandas 对日期时间索引的乘法处理方式不同？

Question

Pandas 乘法对列数据框不起作用:

In [1]: import pandas as pd

In [2]: df = pd.DataFrame({'a': [1,2,3]})

In [3]: s = pd.Series([5,6,7])

In [4]: df * s

    0   1   2   a
0 NaN NaN NaN NaN
1 NaN NaN NaN NaN
2 NaN NaN NaN NaN

如 this question 中所述，正确的做法是specify the axis :

In [5]: df.multiply(s, axis='index')

    a
0   5
1  12
2  21

但是，对于具有日期时间索引的数据帧，这是不必要的:

In [6]: import numpy as np

In [7]: days = np.arange('2000-12-20', '2000-12-23', dtype='datetime64[D]')

In [8]: df = pd.DataFrame({'a': [1,2,3]}, index=days)

In [9]: s = pd.Series([5,6,7], index=days)

In [10]: df * s

             a
2000-12-20   5
2000-12-21  12
2000-12-22  21

这是为什么？为什么第一个示例不起作用，是否可以始终使用 * 而不是 .multiply？

Answer 1

看起来这是较早的弃用行为。问题 here .可能最好在这两种情况下都明确。

In [161]: pd.__version__
Out[161]: '0.16.2'

In [160]: df * s
....: FutureWarning: TimeSeries broadcasting along DataFrame index by default is deprecated. Please use DataFrame.<op> to explicitly broadcast arithmetic operations along the index
  FutureWarning)
Out[160]: 
             a
2000-12-20   5
2000-12-21  12
2000-12-22  21