Python 大熊猫的移位分组序列

我有一个分组序列，我想将其移位并添加到DF中。分组的系列看起来像：

month  day  hour_dep  origin
1      1    5         LGA       1
            6         EWR       4
                      JFK       2
                      LGA       4
            7         EWR       4

当我移动这个序列（test_Series.shift（1））时，我得到

我想要的是：

month  day  hour_dep  origin
1      1    5         LGA       NaN
            6         EWR       NaN
                      JFK       NaN
                      LGA       1.0
            7         EWR       4.0

即，将所有观察值从前一小时转移到下一小时

谢谢你的帮助

一些代码可以轻松复制此数据结构：

np.random.seed(2)
test_df = pd.DataFrame(np.random.randint(0, 5, size=[2000,5]), columns=['Month', 'Day', 'Hour', 'Type', 'Test'])

test_series = test_df.groupby(['Month', 'Day', 'Hour', 'Type']).sum().head(10)['Test']

print(test_series)

print(test_series.shift(1)) # Observations in hour 1 do not correspond to the observations in hour 0. Observations in hour 0 should be NaN.

np.random.seed(2)
test_df = pd.DataFrame(np.random.randint(0, 5, size=[2000,5]), columns=['Month', 'Day', 'Hour', 'Type', 'Test'])

test_series = test_df.groupby(['Month', 'Day', 'Hour', 'Type']).sum().head(10)['Test']

print(test_series)

print(test_series.shift(1)) # Observations in hour 1 do not correspond to the observations in hour 0. Observations in hour 0 should be NaN.