Question

我有一个数据框，df：

                     event
time
2013-10-09 06:27:00   breakfast
2013-10-09 07:25:00      school
2013-10-09 10:25:00        exam
2013-10-09 12:00:00       lunch
2013-10-09 12:00:00    homework
2013-10-09 19:20:00      dinner
2013-10-09 23:25:00       sleep

我还有两个相同大小的时间列表：

lower = array(['2013-10-09T06:15:00.000000000', 
               '2013-10-09T07:00:00.000000000',
               '2013-10-19T13:00:00.000000000',  dtype='datetime64[ns]')

upper = array(['2013-10-09T11:15:00.000000000', 
               '2013-10-09T12:00:00.000000000',
               '2013-10-09T17:00:00.000000000', dtype='datetime64[ns]')

是否有一种非迭代的熊猫式方式可以让我获得df限制的行 [upper [0]，lower [0]]，[upper [1]，lower [1]]，[upper [2]，lower [2]]，...这样我们得到

                     event
time
2013-10-09 06:27:00   breakfast
2013-10-09 07:25:00      school
2013-10-09 10:25:00        exam
2013-10-09 07:25:00      school
2013-10-09 10:25:00        exam
2013-10-09 12:00:00       lunch

作为输出？

Answer 1

一个非常熊猫式的（但可能没有记忆效率）：

(
    events.assign(key=1)
    .merge(pd.DataFrame({"lower": lower, "upper": upper, "key": 1}), on="key")
    [lambda df: df.time.between(d.lower, d.upper)]
    [["time", "event"]]
    .drop_duplicates()
)

Answer 2

不是熊猫的功能，也许你想要什么？

[[u,l] for u, l in zip(upper, lower)]

在日期时间范围列表之间选择pandas行

2 个答案: