Question

我有一些这样的数据：

2013-10-01 11:11:22, 12
2013-10-01 11:15:22, 12
2013-10-01 15:11:22, 12
2013-10-01 16:11:22, 12
2013-10-01 16:45:22, 12

我使用下面的代码重新取样数据：

df.resample('1H', how=lambda x: len(x))

结果如下：

2013-10-01 11:00:00, 2
2013-10-01 12:00:00, 0
2013-10-01 13:00:00, 0
2013-10-01 14:00:00, 0
2013-10-01 15:00:00, 1
2013-10-01 16:00:00, 2

但我希望整天的时间范围从00:00:00到23:00:00，如何更改重新采样日期范围？

Answer 1

嗯，您可以在重新采样后reindex使用一整天的小时段索引。如下所示：

>>> rng = pd.period_range('2013-10-01',periods=24, freq='H').to_timestamp()
>>> df.resample('1H', how=lambda x: len(x)).reindex(index=rng,fill_value=0)

                     value
2013-10-01 00:00:00      0
2013-10-01 01:00:00      0
2013-10-01 02:00:00      0
2013-10-01 03:00:00      0
2013-10-01 04:00:00      0
2013-10-01 05:00:00      0
2013-10-01 06:00:00      0
2013-10-01 07:00:00      0
2013-10-01 08:00:00      0
2013-10-01 09:00:00      0
2013-10-01 10:00:00      0
2013-10-01 11:00:00      2
2013-10-01 12:00:00      0
2013-10-01 13:00:00      0
2013-10-01 14:00:00      0
2013-10-01 15:00:00      1
2013-10-01 16:00:00      2
2013-10-01 17:00:00      0
2013-10-01 18:00:00      0
2013-10-01 19:00:00      0
2013-10-01 20:00:00      0
2013-10-01 21:00:00      0
2013-10-01 22:00:00      0
2013-10-01 23:00:00      0

Answer 2

您应首先使用from datetime import datetime将日期转换为日期时间格式，然后使用pd.Series(datetime.strptime...)。检查https://docs.python.org/2/library/datetime.html，然后您就可以更轻松地操作数据了。

熊猫全天重拍

2 个答案: