COMPAT: datetime parsing with locale

[jreback]

we recently adding testing under a non-US locale for all tests with a larger than 1 unicode character (chinese), here, not just slow (which we have had for a while), and IT (which doesn't show these issues).

#15208 fixes almost everything, but this remains.

You can easily repro this by adding to pandas/__init__.py at the top.

import locale
locale.setlocale(locale.LC_ALL, 'zh_CN.UTF-8')

======================================================================
ERROR: test_to_datetime_format_microsecond (pandas.tseries.tests.test_timeseries.TimeConversionFormats)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/travis/build/jreback/pandas/pandas/tseries/tools.py", line 407, in _convert_listlike
    values, tz = tslib.datetime_to_datetime64(arg)
  File "pandas/tslib.pyx", line 1756, in pandas.tslib.datetime_to_datetime64 (pandas/tslib.c:30671)
    raise TypeError('Unrecognized value type: %s' % type(val))
TypeError: Unrecognized value type: <class 'str'>
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
  File "/home/travis/build/jreback/pandas/pandas/tseries/tests/test_timeseries.py", line 5374, in test_to_datetime_format_microsecond
    result = to_datetime(val, format=format)
  File "/home/travis/build/jreback/pandas/pandas/tseries/tools.py", line 427, in to_datetime
    return _convert_listlike(np.array([arg]), box, format)[0]
  File "/home/travis/build/jreback/pandas/pandas/tseries/tools.py", line 410, in _convert_listlike
    raise e
  File "/home/travis/build/jreback/pandas/pandas/tseries/tools.py", line 377, in _convert_listlike
    errors=errors)
  File "pandas/tslib.pyx", line 3600, in pandas.tslib.array_strptime (pandas/tslib.c:58664)
    raise ValueError("time data %r does not match "
ValueError: time data '01-Apr-2011 00:00:01.978' does not match format '%d-%b-%Y %H:%M:%S.%f' (match)