您可以在不知道格式的情况下使用datetime.strptime吗？

Question

I am writing a function that takes 3 pandas Series, one of which is dates, and I need to be able to turn it into a dataframe where I can resample by them. The issue, is that when I simply do the following:

>>> data.index = data.time
>>> df = data.resample('M')

I get the following error:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/pandas/core/generic.py", line 234, in resample
    return sampler.resample(self)
  File "/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/pandas/tseries/resample.py", line 100, in resample
    raise TypeError('Only valid with DatetimeIndex or PeriodIndex')
TypeError: Only valid with DatetimeIndex or PeriodIndex

I know this is because even though the index type is a datetime object, when going through with resampling, unless it is in the form datetime(x,x,x,x,x,x) , It wont read it correctly.

So when I use it, my date data looks like this: 2011-12-16 08:09:07 , so I have been doing the following:

dates = data.time
date_objects = [datetime.strptime(dates[x], '%Y-%m-%d %H:%M:%S') for x in range(len(dates))]
data.index = date_objects 
df = data.resample('M')

My issue is that I am using this for open source and I cannot know what format the dates will be when inputted.

So my question is: how can I turn a string with a date and a time to a datetime object WITHOUT knowing the way that string is formatted?

Answer 1

You can use the dateutil library for that purpose

from dateutil import parser
yourdate = parser.parse(dates[x])

Answer 2

Pandas has a to_datetime function for this purpose, and when applied to a Series it'll convert values to Timestamp rather than datetime:

data.time = pd.to_datetime(data.time)

df = data.set_index('time')

Where:

In [2]: pd.to_datetime('2011-12-16 08:09:07')
Out[2]: datetime.datetime(2011, 12, 16, 8, 9, 7)

In [3]: s = pd.Series(['2011-12-16 08:09:07'])

In [4]: pd.to_datetime(s)
Out[4]:
0   2011-12-16 08:09:07
dtype: datetime64[ns]