DOC: "Invalid datetime unit in metadata string "[d]""

[jorisvandenbossche]

http://pandas-docs.github.io/pandas-docs-travis/timeseries.html#time-deltas-conversions

In [253]: to_timedelta(np.arange(5),unit='d')
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-253-bab05b276f33> in <module>()
----> 1 to_timedelta(np.arange(5),unit='d')

/home/travis/virtualenv/python2.7_with_system_site_packages/local/lib/python2.7/site-packages/pandas/tseries/timedeltas.pyc in to_timedelta(arg, box, unit)
     61         return Series(values, index=arg.index, name=arg.name, dtype='m8[ns]')
     62     elif is_list_like(arg):
---> 63         return _convert_listlike(arg, box=box, unit=unit)
     64 
     65     # ...so it must be a scalar value. Return scalar.

/home/travis/virtualenv/python2.7_with_system_site_packages/local/lib/python2.7/site-packages/pandas/tseries/timedeltas.pyc in _convert_listlike(arg, box, unit)
     42         elif is_integer_dtype(arg):
     43             # these are shortcutable
---> 44             value = arg.astype('timedelta64[{0}]'.format(unit)).astype('timedelta64[ns]')
     45         else:
     46             try:

TypeError: Invalid datetime unit in metadata string "[d]"

[jorisvandenbossche]

The same error in whatsnew of 0.13: http://pandas-docs.github.io/pandas-docs-travis/whatsnew.html#id11

Did there change something recently?

[jreback]

yeh...i think i made a change, guess I broke this; numpy doesn't like 'd' in timedelta64, only 'D'
(which is odd if you ask me, but ok).