我是我的数据框的子集，但我的情节是相同的

Question

我有返回数据，该数据使用DateTimeIndex索引，如下所示

>ES.head()

Date
2006-01-02 19:15:00   -0.000199
2006-01-02 19:20:00   -0.000199
2006-01-02 19:25:00    0.000000
2006-01-02 19:30:00   -0.000099
2006-01-02 19:35:00   -0.000099
Name: Price, dtype: float64

当我绘制5分钟的平均回报时，我得到以下

fig = plt.figure()
ax1 = fig.add_subplot(111)
plt.plot(ES.groupby([ES.index.hour, ES.index.minute]).mean())
ax1.set_ylabel('Return')
ax1.set_xlabel('5-minute interval')

当我现在在16:00和17:00之间放下所有观察并绘制图像时，我得到相同的图像！？

mask =  ~((ES.index.hour > 16) & (ES.index.hour < 17))
ES = ES[mask]

fig = plt.figure()
ax1 = fig.add_subplot(111)
plt.plot(ES.groupby([ES.index.hour, ES.index.minute]).mean())
ax1.set_ylabel('Return')
ax1.set_xlabel('5-minute interval')

有了这个情节

我已经被困在这个问题上超过一天了，因为我希望在16:00到17:00之间取消由于市场开盘/收盘效应产生的回报。

为什么我的数据框没有正确分组？

编辑：

>ES['2007-01-03 15:30':'2007-01-03 17:30']
Date
2007-01-03 15:30:00    0.000263
2007-01-03 15:35:00    0.000000
2007-01-03 15:40:00   -0.000146
2007-01-03 15:45:00    0.000058
2007-01-03 15:50:00    0.000175
2007-01-03 15:55:00    0.000058
2007-01-03 16:00:00    0.000117
2007-01-03 16:05:00    0.000000
2007-01-03 16:10:00    0.000000
2007-01-03 16:15:00    0.000000
2007-01-03 16:20:00   -0.000175
2007-01-03 16:25:00    0.000175
2007-01-03 16:30:00    0.000000
2007-01-03 16:35:00    0.000000
2007-01-03 16:40:00    0.000000
2007-01-03 16:45:00    0.000000
2007-01-03 16:50:00    0.000000
2007-01-03 16:55:00    0.000000
2007-01-03 17:00:00    0.000058
2007-01-03 17:05:00    0.000292
2007-01-03 17:10:00   -0.000175
2007-01-03 17:15:00    0.000000
2007-01-03 17:20:00    0.000000
2007-01-03 17:25:00    0.000000
2007-01-03 17:30:00    0.000000
Name: Price, dtype: float64

>mask =  ~((ES.index.hour > 16) & (ES.index.hour < 17))
>ES = ES[mask]

>ES['2007-01-03 15:30':'2007-01-03 17:30']
Date
2007-01-03 15:30:00    0.000263
2007-01-03 15:35:00    0.000000
2007-01-03 15:40:00   -0.000146
2007-01-03 15:45:00    0.000058
2007-01-03 15:50:00    0.000175
2007-01-03 15:55:00    0.000058
2007-01-03 16:00:00    0.000117
2007-01-03 16:05:00    0.000000
2007-01-03 16:10:00    0.000000
2007-01-03 16:15:00    0.000000
2007-01-03 16:20:00   -0.000175
2007-01-03 16:25:00    0.000175
2007-01-03 16:30:00    0.000000
2007-01-03 16:35:00    0.000000
2007-01-03 16:40:00    0.000000
2007-01-03 16:45:00    0.000000
2007-01-03 16:50:00    0.000000
2007-01-03 16:55:00    0.000000
2007-01-03 17:00:00    0.000058
2007-01-03 17:05:00    0.000292
2007-01-03 17:10:00   -0.000175
2007-01-03 17:15:00    0.000000
2007-01-03 17:20:00    0.000000
2007-01-03 17:25:00    0.000000
2007-01-03 17:30:00    0.000000
Name: Price, dtype: float64

似乎我的面具不起作用，但是，当我尝试以下时，在删除适当的行时，绘图仍然相同。

>ES['2007-01-03 15:30':'2007-01-03 17:30']
Date
2007-01-03 15:30:00    0.000263
2007-01-03 15:35:00    0.000000
2007-01-03 15:40:00   -0.000146
2007-01-03 15:45:00    0.000058
2007-01-03 15:50:00    0.000175
2007-01-03 15:55:00    0.000058
2007-01-03 16:00:00    0.000117
2007-01-03 16:05:00    0.000000
2007-01-03 16:10:00    0.000000
2007-01-03 16:15:00    0.000000
2007-01-03 16:20:00   -0.000175
2007-01-03 16:25:00    0.000175
2007-01-03 16:30:00    0.000000
2007-01-03 16:35:00    0.000000
2007-01-03 16:40:00    0.000000
2007-01-03 16:45:00    0.000000
2007-01-03 16:50:00    0.000000
2007-01-03 16:55:00    0.000000
2007-01-03 17:00:00    0.000058
2007-01-03 17:05:00    0.000292
2007-01-03 17:10:00   -0.000175
2007-01-03 17:15:00    0.000000
2007-01-03 17:20:00    0.000000
2007-01-03 17:25:00    0.000000
2007-01-03 17:30:00    0.000000
Name: Price, dtype: float64
    ​
>ES = ES.drop(ES.between_time('16:00','17:00').index)

>ES['2007-01-03 15:30':'2007-01-03 17:30']
Date
2007-01-03 15:30:00    0.000263
2007-01-03 15:35:00    0.000000
2007-01-03 15:40:00   -0.000146
2007-01-03 15:45:00    0.000058
2007-01-03 15:50:00    0.000175
2007-01-03 15:55:00    0.000058
2007-01-03 17:05:00    0.000292
2007-01-03 17:10:00   -0.000175
2007-01-03 17:15:00    0.000000
2007-01-03 17:20:00    0.000000
2007-01-03 17:25:00    0.000000
2007-01-03 17:30:00    0.000000
Name: Price, dtype: float64