.loc interprets slice as positional rather than label-based when setting

[toobaz]

Code Sample, a copy-pastable example if possible

In [2]: df = pd.DataFrame(-1, index=['i', 'ii', 'iii'], columns=pd.MultiIndex.from_tuples([['A', 'a'], ['B', 'b']]))

In [3]: df.loc[1:, 'A'] = ''

In [4]: df
Out[4]: 
      A  B
      a  b
i    -1 -1
ii      -1
iii     -1

In [5]: df.loc[1:, 'A']
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-5-98b6153a5b09> in <module>()
----> 1 df.loc[1:, 'A']

/home/pietro/nobackup/repo/pandas/pandas/core/indexing.py in __getitem__(self, key)
   1322             except (KeyError, IndexError):
    948 
    949         # we maybe be using a tuple to represent multiple dimensions here

/home/pietro/nobackup/repo/pandas/pandas/core/indexing.py in _getitem_nested_tuple(self, tup)
   1020 
   1021             current_ndim = obj.ndim
-> 1022             obj = getattr(obj, self.name)._getitem_axis(key, axis=axis)
   1023             axis += 1
   1024 

/home/pietro/nobackup/repo/pandas/pandas/core/indexing.py in _getitem_axis(self, key, axis)
   1503         if isinstance(key, slice):
   1504             self._has_valid_type(key, axis)
-> 1505             return self._get_slice_axis(key, axis=axis)
   1506         elif is_bool_indexer(key):
   1507             return self._getbool_axis(key, axis=axis)

/home/pietro/nobackup/repo/pandas/pandas/core/indexing.py in _get_slice_axis(self, slice_obj, axis)
   1353         labels = obj._get_axis(axis)
   1354         indexer = labels.slice_indexer(slice_obj.start, slice_obj.stop,
-> 1355                                        slice_obj.step, kind=self.name)
   1356 
   1357         if isinstance(indexer, slice):

/home/pietro/nobackup/repo/pandas/pandas/indexes/base.py in slice_indexer(self, start, end, step, kind)
   3247         """
   3248         start_slice, end_slice = self.slice_locs(start, end, step=step,
-> 3249                                                  kind=kind)
   3250 
   3251         # return a slice

/home/pietro/nobackup/repo/pandas/pandas/indexes/base.py in slice_locs(self, start, end, step, kind)
   3428         start_slice = None
   3429         if start is not None:
-> 3430             start_slice = self.get_slice_bound(start, 'left', kind)
   3431         if start_slice is None:
   3432             start_slice = 0

/home/pietro/nobackup/repo/pandas/pandas/indexes/base.py in get_slice_bound(self, label, side, kind)
   3367         # For datetime indices label may be a string that has to be converted
   3368         # to datetime boundary according to its resolution.
-> 3369         label = self._maybe_cast_slice_bound(label, side, kind)
   3370 
   3371         # we need to look up the label

/home/pietro/nobackup/repo/pandas/pandas/indexes/base.py in _maybe_cast_slice_bound(self, label, side, kind)
   3325         # this is rejected (generally .loc gets you here)
   3326         elif is_integer(label):
-> 3327             self._invalid_indexer('slice', label)
   3328 
   3329         return label

/home/pietro/nobackup/repo/pandas/pandas/indexes/base.py in _invalid_indexer(self, form, key)
   1447                         "indexers [{key}] of {kind}".format(
   1448                             form=form, klass=type(self), key=key,
-> 1449                             kind=type(key)))
   1450 
   1451     def get_duplicates(self):

TypeError: cannot do slice indexing on <class 'pandas.indexes.base.Index'> with these indexers [1] of <class 'int'>

Problem description

In[3]: should already raise an error (unless the label 1 is compared against the labels in the index, in which case it should certainly be smaller or larger than each of them).

Expected Output

An error

Output of pd.show_versions()

INSTALLED VERSIONS

commit: None
python: 3.5.3.final.0
python-bits: 64
OS: Linux
OS-release: 3.16.0-4-amd64
machine: x86_64
processor:
byteorder: little
LC_ALL: None
LANG: en_GB.UTF-8
LOCALE: en_GB.UTF-8

pandas: 0.19.0+783.gcd35d22a0
pytest: 3.0.6
pip: 1.5.6
setuptools: 5.5.1
Cython: 0.25.2
numpy: 1.12.1
scipy: 0.18.1
xarray: None
IPython: 5.2.2
sphinx: None
patsy: 0.4.1+dev
dateutil: 2.6.0
pytz: 2016.10
blosc: None
bottleneck: None
tables: 3.3.0
numexpr: 2.6.1
feather: 0.3.1
matplotlib: 2.0.0
openpyxl: 2.3.0
xlrd: 0.9.2
xlwt: 1.2.0
xlsxwriter: None
lxml: 3.7.1
bs4: None
html5lib: 0.999
sqlalchemy: 0.9.8
pymysql: None
psycopg2: None
jinja2: 2.7.3
s3fs: None
pandas_gbq: None
pandas_datareader: None

[jorisvandenbossche]

Note that in this case the MultiIndex is not needed to reproduce the issue (which you didn't say, but just to not complicate the example):

In [27]: df = pd.DataFrame(-1, index=['i', 'ii', 'iii'], columns=['A', 'B'])

In [28]: df.loc[1:, 'A']
...
TypeError: cannot do slice indexing on <class 'pandas.core.indexes.base.Index'> with these indexers [1] of <class 'int'>

In [29]: df.loc[1:, 'A'] = 1

In [30]: df
Out[30]: 
     A  B
i   -1 -1
ii   1 -1
iii  1 -1