python - 根据其他单元格值更改多个单元格值

Question

我想根据下一个单元格中给出的状态，将包含 moving 的单元格更改为 movingToOpen 或 movingToClose。

有时循环会中断，不会从open 移动到close 或从close 移动到open

这是我当前的数据框:

                   DateTime  moving
id                                 
36  2018-11-12 15:06:02.487    open
37  2018-11-12 15:06:03.691  moving
38  2018-11-12 15:06:05.505  moving
39  2018-11-12 15:06:08.622   close
40  2018-11-12 15:06:09.023   close
41  2018-11-12 15:06:09.931   close
42  2018-11-12 15:06:11.944  moving
43  2018-11-12 15:06:13.756  moving
44  2018-11-12 15:06:15.168  moving
45  2018-11-12 15:06:18.388   close
46  2018-11-12 15:06:20.100  moving
47  2018-11-12 15:06:23.316  moving
48  2018-11-12 15:06:25.730    open
49  2018-11-12 15:06:26.637    open
50  2018-11-12 15:06:27.644    open
51  2018-11-12 15:06:28.550    open
52  2018-11-12 15:06:28.855    open
53  2018-11-12 15:06:29.356    open
54  2018-11-12 15:06:30.563    open
55  2018-11-12 15:06:31.369  moving
56  2018-11-12 15:06:32.575  moving
57  2018-11-12 15:06:35.593   close

我希望它看起来像这样:

                   DateTime  moving
id                                 
36  2018-11-12 15:06:02.487    open
37  2018-11-12 15:06:03.691  movingToClose
38  2018-11-12 15:06:05.505  movingToClose
39  2018-11-12 15:06:08.622   close
40  2018-11-12 15:06:09.023   close
41  2018-11-12 15:06:09.931   close
42  2018-11-12 15:06:11.944  movingToClose
43  2018-11-12 15:06:13.756  movingToClose
44  2018-11-12 15:06:15.168  movingToClose
45  2018-11-12 15:06:18.388   close
46  2018-11-12 15:06:20.100  movingToOpen
47  2018-11-12 15:06:23.316  movingToOpen
48  2018-11-12 15:06:25.730    open
49  2018-11-12 15:06:26.637    open
50  2018-11-12 15:06:27.644    open
51  2018-11-12 15:06:28.550    open
52  2018-11-12 15:06:28.855    open
53  2018-11-12 15:06:29.356    open
54  2018-11-12 15:06:30.563    open
55  2018-11-12 15:06:31.369  movingToClose
56  2018-11-12 15:06:32.575  movingToClose
57  2018-11-12 15:06:35.593   close

我已经尝试过将其转换为数字并使用 np.ediff1d 的方法，但这只会让它发生在一行中，而且在我看来有一种更简单的方法。

Answer 1

首先使用 mask 将 moving 值转换为 NaN ，回填最后一个非缺失值并添加到掩码过滤的原始值:

m = df['moving'] == 'moving'
df.loc[m, 'moving'] = 'movingTo' + df['moving'].mask(m).bfill()
print (df)
                   DateTime         moving
36  2018-11-12 15:06:02.487           open
37  2018-11-12 15:06:03.691  movingToclose
38  2018-11-12 15:06:05.505  movingToclose
39  2018-11-12 15:06:08.622          close
40  2018-11-12 15:06:09.023          close
41  2018-11-12 15:06:09.931          close
42  2018-11-12 15:06:11.944  movingToclose
43  2018-11-12 15:06:13.756  movingToclose
44  2018-11-12 15:06:15.168  movingToclose
45  2018-11-12 15:06:18.388          close
46  2018-11-12 15:06:20.100   movingToopen
47  2018-11-12 15:06:23.316   movingToopen
48  2018-11-12 15:06:25.730           open
49  2018-11-12 15:06:26.637           open
50  2018-11-12 15:06:27.644           open
51  2018-11-12 15:06:28.550           open
52  2018-11-12 15:06:28.855           open
53  2018-11-12 15:06:29.356           open
54  2018-11-12 15:06:30.563           open
55  2018-11-12 15:06:31.369  movingToclose
56  2018-11-12 15:06:32.575  movingToclose
57  2018-11-12 15:06:35.593          close