BUG: .drop() doesn't parse strings for DatetimeIndex.

[jtratner]

Needs to coerce:

e.g.

In [1]: import pandas.util.testing as tm

In [2]: pan = tm.makePanel()

In [3]: pan
Out[3]:
<class 'pandas.core.panel.Panel'>
Dimensions: 3 (items) x 30 (major_axis) x 4 (minor_axis)
Items axis: ItemA to ItemC
Major_axis axis: 2000-01-03 00:00:00 to 2000-02-11 00:00:00
Minor_axis axis: A to D

In [4]: pan.drop('2000-01-03', axis=1)
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-4-11e47d7c99f1> in <module>()
----> 1 pan.drop('2000-01-03', axis=1)

../pandas/core/generic.pyc in drop(self, labels, axis, level)
   1112                 new_axis = axis.drop(labels, level=level)
   1113             else:
-> 1114                 new_axis = axis.drop(labels)
   1115             dropped = self.reindex(**{ axis_name: new_axis })
   1116             try:

../pandas/core/index.pyc in drop(self, labels)
   1614         mask = indexer == -1
   1615         if mask.any():
-> 1616             raise ValueError('labels %s not contained in axis' % labels[mask])
   1617         return self.delete(indexer)
   1618

ValueError: labels ['2000-01-03'] not contained in axis

[jreback]

related #6599

[jorisvandenbossche]

Updated example with frame:

In [59]: df = tm.makeTimeDataFrame()

In [60]: df
Out[60]: 
                   A         B         C         D
2000-01-03 -0.394676  0.026933  1.435611  0.735617
2000-01-04 -2.095283 -0.623525  1.017788 -2.014091
2000-01-05 -0.770375  0.643865 -0.070327 -2.305731
2000-01-06 -1.188628  0.475190 -0.926808 -0.285167
2000-01-07 -0.125856  1.719509  0.184650  0.640200
...

In [62]: df.drop('2000-01-03', axis=0)
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-62-0844c2a2456f> in <module>()
----> 1 df.drop('2000-01-03', axis=0)

/home/joris/scipy/pandas/pandas/core/generic.py in drop(self, labels, axis, index, columns, level, inplace, errors)
   2515         for axis, labels in axes.items():
   2516             if labels is not None:
-> 2517                 obj = obj._drop_axis(labels, axis, level=level, errors=errors)
   2518 
   2519         if inplace:

/home/joris/scipy/pandas/pandas/core/generic.py in _drop_axis(self, labels, axis, level, errors)
   2547                 new_axis = axis.drop(labels, level=level, errors=errors)
   2548             else:
-> 2549                 new_axis = axis.drop(labels, errors=errors)
   2550             dropped = self.reindex(**{axis_name: new_axis})
   2551             try:

/home/joris/scipy/pandas/pandas/core/indexes/base.py in drop(self, labels, errors)
   3757             if errors != 'ignore':
   3758                 raise ValueError('labels %s not contained in axis' %
-> 3759                                  labels[mask])
   3760             indexer = indexer[~mask]
   3761         return self.delete(indexer)

ValueError: labels ['2000-01-03'] not contained in axis

In [63]: df.drop(pd.Timestamp('2000-01-03'), axis=0)
Out[63]: 
                   A         B         C         D
2000-01-04 -3.354600 -0.610585 -0.503516 -1.058210
2000-01-05  2.253061  0.083150  0.236107 -1.475713
2000-01-06 -0.299333  0.957344  2.024540 -0.424540
2000-01-07  0.396877 -1.124760 -0.249968 -1.139085
...

[jreback]

@jorisvandenbossche why do you think this should work? this is partial string indexing. I mean it could, but I think drop being strict is a good thing.

[jorisvandenbossche]

I am not necessarily saying that this should work. I was going through old issues and closing/updating some, so updated the title / added a better example of the original issue here.

I agree partial strings interpretation should not work with drop, but in the example above, it is not a partial string that embodies multiple rows. It is just a single label represented by a string instead of Timestamp object. So supporting that is I think a bit less controversial.

[mroeschke]

Looks to work on main now. Could use a test