在特定日期和时间之前/之后截断错误

Question

我编写了一个函数，用于在给定范围之前或之后截断数据。我传入一个日期元组，位置1是开始日期，位置2是结束日期。

我如何也可以选择指定截断b4和b4之后的时间和日期-我将如何使用我的代码来做到这一点？*）

此外-我一直在我的数据上得到一个错误，说：

    raise InvalidIndexError('Reindexing only valid with uniquely'
pandas.core.indexes.base.InvalidIndexError: Reindexing only valid with uniquely valued Index
objects

我使用：

date_range = self.data.copy()
nearest_start = date_range.iloc[
            date_range.index.get_loc(datetime.datetime.strptime(split[0], '%m-%d-%Y'),
                     method='nearest')].name
...
date_range = date_range.truncate(before=nearest_start, after=nearest_end)

我的日期元组：

('12-29-2011','12-23-2017')

这是数据集的my df.head（）：

                               Open       H   ...                RBE1min        O
        DateTime                              ...                                
        2007-11-06 12:45:00 -0.6437 -0.6423   ...     11/6/2007 12:45.00  21198.0
        2007-11-06 14:30:00 -0.6430 -0.6425   ...     11/6/2007 14:30.00  21355.0
        2007-11-06 19:33:00 -0.6423 -0.6400   ...     11/6/2007 19:33.00  21430.0
        2007-11-07 09:00:00 -0.6434 -0.6421   ...     11/7/2007 09:00.00  21440.0
        2007-11-07 12:26:00 -0.6433 -0.6396   ...     11/7/2007 12:26.00  21392.0

我已经运行了下面的程序来检查索引中的dupilcates，它返回NAT，所以我不确定为什么我仍然得到这个错误

任何帮助都将不胜感激

dup = rbs4181.index.get_duplicates()

Answer 1

恐怕不是你想要的。但这可能会有所帮助：

我使用此数据集http://archive.ics.uci.edu/ml/datasets/Occupancy检测#进行测试。

对于datetime索引，解析为时间戳的日期和字符串可以作为索引参数传递，也可以作为truncate和pd的参数之前和之后的传递。时间戳.

In [1]: import pandas as pd

In [2]: df = pd.read_csv('datatest2.txt', parse_dates=[1], index_col=[1])

In [3]: df.index
Out[3]:
DatetimeIndex(['2015-02-11 14:48:00', '2015-02-11 14:49:00',
               '2015-02-11 14:50:00', '2015-02-11 14:51:00',
               '2015-02-11 14:51:59', '2015-02-11 14:53:00',
               '2015-02-11 14:54:00', '2015-02-11 14:55:00',
               '2015-02-11 14:55:59', '2015-02-11 14:57:00',
               ...
               '2015-02-18 09:10:00', '2015-02-18 09:10:59',
               '2015-02-18 09:11:59', '2015-02-18 09:13:00',
               '2015-02-18 09:14:00', '2015-02-18 09:15:00',
               '2015-02-18 09:16:00', '2015-02-18 09:16:59',
               '2015-02-18 09:17:59', '2015-02-18 09:19:00'],
              dtype='datetime64[ns]', name='date', length=9752, freq=None)

In [4]: df['2015-02-12':'2015-02-13'].index
Out[4]:
DatetimeIndex(['2015-02-12 00:00:00', '2015-02-12 00:01:00',
               '2015-02-12 00:02:00', '2015-02-12 00:03:00',
               '2015-02-12 00:04:00', '2015-02-12 00:04:59',
               '2015-02-12 00:06:00', '2015-02-12 00:07:00',
               '2015-02-12 00:08:00', '2015-02-12 00:08:59',
               ...
               '2015-02-13 23:50:00', '2015-02-13 23:51:00',
               '2015-02-13 23:51:59', '2015-02-13 23:53:00',
               '2015-02-13 23:54:00', '2015-02-13 23:55:00',
               '2015-02-13 23:55:59', '2015-02-13 23:57:00',
               '2015-02-13 23:57:59', '2015-02-13 23:58:59'],
              dtype='datetime64[ns]', name='date', length=2880, freq=None)

In [5]: df.truncate(before=pd.Timestamp('2015-02-12'), after=pd.Timestamp('2015-02-14')).index
Out[5]:
DatetimeIndex(['2015-02-12 00:00:00', '2015-02-12 00:01:00',
               '2015-02-12 00:02:00', '2015-02-12 00:03:00',
               '2015-02-12 00:04:00', '2015-02-12 00:04:59',
               '2015-02-12 00:06:00', '2015-02-12 00:07:00',
               '2015-02-12 00:08:00', '2015-02-12 00:08:59',
               ...
               '2015-02-13 23:51:00', '2015-02-13 23:51:59',
               '2015-02-13 23:53:00', '2015-02-13 23:54:00',
               '2015-02-13 23:55:00', '2015-02-13 23:55:59',
               '2015-02-13 23:57:00', '2015-02-13 23:57:59',
               '2015-02-13 23:58:59', '2015-02-14 00:00:00'],
              dtype='datetime64[ns]', name='date', length=2881, freq=None)

In [6]: df.truncate(before='2015-02-12', after='2015-02-14').index
Out[6]:
DatetimeIndex(['2015-02-12 00:00:00', '2015-02-12 00:01:00',
               '2015-02-12 00:02:00', '2015-02-12 00:03:00',
               '2015-02-12 00:04:00', '2015-02-12 00:04:59',
               '2015-02-12 00:06:00', '2015-02-12 00:07:00',
               '2015-02-12 00:08:00', '2015-02-12 00:08:59',
               ...
               '2015-02-13 23:51:00', '2015-02-13 23:51:59',
               '2015-02-13 23:53:00', '2015-02-13 23:54:00',
               '2015-02-13 23:55:00', '2015-02-13 23:55:59',
               '2015-02-13 23:57:00', '2015-02-13 23:57:59',
               '2015-02-13 23:58:59', '2015-02-14 00:00:00'],
              dtype='datetime64[ns]', name='date', length=2881, freq=None)

In [7]: df.truncate(before='2015-02-12 01:00:00', after='2015-02-13 23:00:00').index
Out[7]:
DatetimeIndex(['2015-02-12 01:00:00', '2015-02-12 01:01:00',
               '2015-02-12 01:01:59', '2015-02-12 01:02:59',
               '2015-02-12 01:04:00', '2015-02-12 01:05:00',
               '2015-02-12 01:06:00', '2015-02-12 01:07:00',
               '2015-02-12 01:08:00', '2015-02-12 01:08:59',
               ...
               '2015-02-13 22:51:00', '2015-02-13 22:52:00',
               '2015-02-13 22:53:00', '2015-02-13 22:53:59',
               '2015-02-13 22:54:59', '2015-02-13 22:56:00',
               '2015-02-13 22:57:00', '2015-02-13 22:58:00',
               '2015-02-13 22:59:00', '2015-02-13 22:59:59'],
              dtype='datetime64[ns]', name='date', length=2761, freq=None)

因此，我认为您只需要修改函数来验证是否传递了有效日期（和时间）。