使用inferred_freq设置pandas.tseries.index.DatetimeIndex.freq

Question

考虑 DatetimeIndex tidx

tidx = pd.to_datetime(['2016-07-29', '2016-08-31', '2016-09-30'])
print(tidx.freq)
print(tidx.inferred_freq)
print(tidx)

None
BM
DatetimeIndex(['2016-07-29', '2016-08-31', '2016-09-30'], \
              dtype='datetime64[ns]', freq=None)

我希望freq属性采用inferred_freq属性…所以我

tidx.freq = tidx.inferred_freq
print(tidx)

AttributeError Traceback (most recent call last)
in ()
1 tidx.freq = tidx.inferred_freq
----> 2 print(tidx)

C:\Users\ssmith\AppData\Local\Continuum\Anaconda2\lib\site-

packages\pandas\core\base.pyc in str(self)
45 if compat.PY3:
46 return self.unicode()
—> 47 return self.bytes()
48
49 def bytes(self):

C:\Users\ssmith\AppData\Local\Continuum\Anaconda2\lib\site-

packages\pandas\core\base.pyc in bytes(self)
57
58 encoding = get_option(“display.encoding”)
—> 59 return self.unicode().encode(encoding, ‘replace’)
60
61 def repr(self):

C:\Users\ssmith\AppData\Local\Continuum\Anaconda2\lib\site-

packages\pandas\indexes\base.pyc in unicode(self)
581 klass = self.class.name
582 data = self._format_data()
–> 583 attrs = self._format_attrs()
584 space = self._format_space()
585

C:\Users\ssmith\AppData\Local\Continuum\Anaconda2\lib\site-

packages\pandas\tseries\base.pyc in _format_attrs(self)
485 for attrib in self._attributes:
486 if attrib == ‘freq’:
–> 487 freq = self.freqstr
488 if freq is not None:
489 freq = “’%s’” % freq

C:\Users\ssmith\AppData\Local\Continuum\Anaconda2\lib\site-

packages\pandas\tseries\base.pyc in freqstr(self)
223 if self.freq is None:
224 return None
–> 225 return self.freq.freqstr
226
227 @cache_readonly

AttributeError: 'str' object has no attribute 'freqstr'

问题
使索引假定为inferred_freq的合适方法是什么？

Answer 1

目前尚不清楚为什么文档说明您可以设置该freq属性，但是该属性不能持久存在，但是如果您datetimeindex再次freq构造该属性但传递了一个参数，那么它将起作用：

In [56]:
tidx = pd.DatetimeIndex(tidx.values, freq = tidx.inferred_freq)
tidx

Out[56]:
DatetimeIndex(['2016-07-29', '2016-08-31', '2016-09-30'], dtype='datetime64[ns]', freq='BM')