indexing.py

# pylint: disable=W0223
from copy import deepcopy
import textwrap
import warnings
import numpy as np
from pandas.compat import range, zip
import pandas.compat as compat
from pandas.core.dtypes.generic import ABCDataFrame, ABCPanel, ABCSeries
from pandas.core.dtypes.common import (
    is_integer_dtype,
    is_integer, is_float,
    is_list_like,
    is_sequence,
    is_iterator,
    is_scalar,
    is_sparse,
    _ensure_platform_int)
from pandas.core.dtypes.missing import isna, _infer_fill_value
from pandas.errors import AbstractMethodError
from pandas.util._decorators import Appender

from pandas.core.index import Index, MultiIndex

import pandas.core.common as com
from pandas._libs.indexing import _NDFrameIndexerBase


# the supported indexers
def get_indexers_list():

    return [
        ('ix', _IXIndexer),
        ('iloc', _iLocIndexer),
        ('loc', _LocIndexer),
        ('at', _AtIndexer),
        ('iat', _iAtIndexer),
    ]


# "null slice"
_NS = slice(None, None)


# the public IndexSlicerMaker
class _IndexSlice(object):
    """
    Create an object to more easily perform multi-index slicing

    See Also
    --------
    MultiIndex.remove_unused_levels : New MultiIndex with no unused levels.

    Notes
    -----
    See :ref:`Defined Levels <advanced.shown_levels>`
    for further info on slicing a MultiIndex.

    Examples
    --------

    >>> midx = pd.MultiIndex.from_product([['A0','A1'], ['B0','B1','B2','B3']])
    >>> columns = ['foo', 'bar']
    >>> dfmi = pd.DataFrame(np.arange(16).reshape((len(midx), len(columns))),
                            index=midx, columns=columns)

    Using the default slice command:

    >>> dfmi.loc[(slice(None), slice('B0', 'B1')), :]
               foo  bar
        A0 B0    0    1
           B1    2    3
        A1 B0    8    9
           B1   10   11

    Using the IndexSlice class for a more intuitive command:

    >>> idx = pd.IndexSlice
    >>> dfmi.loc[idx[:, 'B0':'B1'], :]
               foo  bar
        A0 B0    0    1
           B1    2    3
        A1 B0    8    9
           B1   10   11
    """

    def __getitem__(self, arg):
        return arg


IndexSlice = _IndexSlice()


class IndexingError(Exception):
    pass


class _NDFrameIndexer(_NDFrameIndexerBase):
    _valid_types = None
    _exception = KeyError
    axis = None

    def __call__(self, axis=None):
        # we need to return a copy of ourselves
        new_self = self.__class__(self.name, self.obj)

        if axis is not None:
            axis = self.obj._get_axis_number(axis)
        new_self.axis = axis
        return new_self

    def __iter__(self):
        raise NotImplementedError('ix is not iterable')

    def __getitem__(self, key):
        if type(key) is tuple:
            key = tuple(com._apply_if_callable(x, self.obj)
                        for x in key)
            try:
                values = self.obj._get_value(*key)
                if is_scalar(values):
                    return values
            except Exception:
                pass

            return self._getitem_tuple(key)
        else:
            # we by definition only have the 0th axis
            axis = self.axis or 0

            key = com._apply_if_callable(key, self.obj)
            return self._getitem_axis(key, axis=axis)

    def _get_label(self, label, axis=None):
        if axis is None:
            axis = self.axis or 0

        if self.ndim == 1:
            # for perf reasons we want to try _xs first
            # as its basically direct indexing
            # but will fail when the index is not present
            # see GH5667
            return self.obj._xs(label, axis=axis)
        elif isinstance(label, tuple) and isinstance(label[axis], slice):
            raise IndexingError('no slices here, handle elsewhere')

        return self.obj._xs(label, axis=axis)

    def _get_loc(self, key, axis=None):
        if axis is None:
            axis = self.axis
        return self.obj._ixs(key, axis=axis)

    def _slice(self, obj, axis=None, kind=None):
        if axis is None:
            axis = self.axis
        return self.obj._slice(obj, axis=axis, kind=kind)

    def _get_setitem_indexer(self, key):
        if self.axis is not None:
            return self._convert_tuple(key, is_setter=True)

        axis = self.obj._get_axis(0)

        if isinstance(axis, MultiIndex) and self.name != 'iloc':
            try:
                return axis.get_loc(key)
            except Exception:
                pass

        if isinstance(key, tuple):
            try:
                return self._convert_tuple(key, is_setter=True)
            except IndexingError:
                pass

        if isinstance(key, range):
            return self._convert_range(key, is_setter=True)

        try:
            return self._convert_to_indexer(key, is_setter=True)
        except TypeError as e:

            # invalid indexer type vs 'other' indexing errors
            if 'cannot do' in str(e):
                raise
            raise IndexingError(key)

    def __setitem__(self, key, value):
        if isinstance(key, tuple):
            key = tuple(com._apply_if_callable(x, self.obj)
                        for x in key)
        else:
            key = com._apply_if_callable(key, self.obj)
        indexer = self._get_setitem_indexer(key)
        self._setitem_with_indexer(indexer, value)

    def _validate_key(self, key, axis):
        """
        Ensure that key is valid for current indexer.

        Parameters
        ----------
        key : scalar, slice or list-like
            The key requested

        axis : int
            Dimension on which the indexing is being made

        Raises
        ------
        TypeError
            If the key (or some element of it) has wrong type

        IndexError
            If the key (or some element of it) is out of bounds

        KeyError
            If the key was not found
        """
        raise AbstractMethodError()

    def _has_valid_tuple(self, key):
        """ check the key for valid keys across my indexer """
        for i, k in enumerate(key):
            if i >= self.obj.ndim:
                raise IndexingError('Too many indexers')
            try:
                self._validate_key(k, i)
            except ValueError:
                raise ValueError("Location based indexing can only have "
                                 "[{types}] types"
                                 .format(types=self._valid_types))

    def _is_nested_tuple_indexer(self, tup):
        if any(isinstance(ax, MultiIndex) for ax in self.obj.axes):
            return any(is_nested_tuple(tup, ax) for ax in self.obj.axes)
        return False

    def _convert_tuple(self, key, is_setter=False):
        keyidx = []
        if self.axis is not None:
            axis = self.obj._get_axis_number(self.axis)
            for i in range(self.ndim):
                if i == axis:
                    keyidx.append(self._convert_to_indexer(
                        key, axis=axis, is_setter=is_setter))
                else:
                    keyidx.append(slice(None))
        else:
            for i, k in enumerate(key):
                if i >= self.obj.ndim:
                    raise IndexingError('Too many indexers')
                idx = self._convert_to_indexer(k, axis=i, is_setter=is_setter)
                keyidx.append(idx)
        return tuple(keyidx)

    def _convert_range(self, key, is_setter=False):
        """ convert a range argument """
        return list(key)

    def _convert_scalar_indexer(self, key, axis):
        # if we are accessing via lowered dim, use the last dim
        if axis is None:
            axis = 0
        ax = self.obj._get_axis(min(axis, self.ndim - 1))
        # a scalar
        return ax._convert_scalar_indexer(key, kind=self.name)

    def _convert_slice_indexer(self, key, axis):
        # if we are accessing via lowered dim, use the last dim
        ax = self.obj._get_axis(min(axis, self.ndim - 1))
        return ax._convert_slice_indexer(key, kind=self.name)

    def _has_valid_setitem_indexer(self, indexer):
        return True

    def _has_valid_positional_setitem_indexer(self, indexer):
        """ validate that an positional indexer cannot enlarge its target
        will raise if needed, does not modify the indexer externally
        """
        if isinstance(indexer, dict):
            raise IndexError("{0} cannot enlarge its target object"
                             .format(self.name))
        else:
            if not isinstance(indexer, tuple):
                indexer = self._tuplify(indexer)
            for ax, i in zip(self.obj.axes, indexer):
                if isinstance(i, slice):
                    # should check the stop slice?
                    pass
                elif is_list_like_indexer(i):
                    # should check the elements?
                    pass
                elif is_integer(i):
                    if i >= len(ax):
                        raise IndexError("{name} cannot enlarge its target "
                                         "object".format(name=self.name))
                elif isinstance(i, dict):
                    raise IndexError("{name} cannot enlarge its target object"
                                     .format(name=self.name))

        return True

    def _setitem_with_indexer(self, indexer, value):
        self._has_valid_setitem_indexer(indexer)

        # also has the side effect of consolidating in-place
        # TODO: Panel, DataFrame are not imported, remove?
        from pandas import Panel, DataFrame, Series  # noqa
        info_axis = self.obj._info_axis_number

        # maybe partial set
        take_split_path = self.obj._is_mixed_type

        # if there is only one block/type, still have to take split path
        # unless the block is one-dimensional or it can hold the value
        if not take_split_path and self.obj._data.blocks:
            blk, = self.obj._data.blocks
            if 1 < blk.ndim:  # in case of dict, keys are indices
                val = list(value.values()) if isinstance(value,
                                                         dict) else value
                take_split_path = not blk._can_hold_element(val)

        if isinstance(indexer, tuple) and len(indexer) == len(self.obj.axes):

            for i, ax in zip(indexer, self.obj.axes):

                # if we have any multi-indexes that have non-trivial slices
                # (not null slices) then we must take the split path, xref
                # GH 10360
                if (isinstance(ax, MultiIndex) and
                        not (is_integer(i) or com.is_null_slice(i))):
                    take_split_path = True
                    break

        if isinstance(indexer, tuple):
            nindexer = []
            for i, idx in enumerate(indexer):
                if isinstance(idx, dict):

                    # reindex the axis to the new value
                    # and set inplace
                    key, _ = convert_missing_indexer(idx)

                    # if this is the items axes, then take the main missing
                    # path first
                    # this correctly sets the dtype and avoids cache issues
                    # essentially this separates out the block that is needed
                    # to possibly be modified
                    if self.ndim > 1 and i == self.obj._info_axis_number:

                        # add the new item, and set the value
                        # must have all defined axes if we have a scalar
                        # or a list-like on the non-info axes if we have a
                        # list-like
                        len_non_info_axes = [
                            len(_ax) for _i, _ax in enumerate(self.obj.axes)
                            if _i != i
                        ]
                        if any(not l for l in len_non_info_axes):
                            if not is_list_like_indexer(value):
                                raise ValueError("cannot set a frame with no "
                                                 "defined index and a scalar")
                            self.obj[key] = value
                            return self.obj

                        # add a new item with the dtype setup
                        self.obj[key] = _infer_fill_value(value)

                        new_indexer = convert_from_missing_indexer_tuple(
                            indexer, self.obj.axes)
                        self._setitem_with_indexer(new_indexer, value)

                        return self.obj

                    # reindex the axis
                    # make sure to clear the cache because we are
                    # just replacing the block manager here
                    # so the object is the same
                    index = self.obj._get_axis(i)
                    labels = index.insert(len(index), key)
                    self.obj._data = self.obj.reindex(labels, axis=i)._data
                    self.obj._maybe_update_cacher(clear=True)
                    self.obj._is_copy = None

                    nindexer.append(labels.get_loc(key))

                else:
                    nindexer.append(idx)

            indexer = tuple(nindexer)
        else:

            indexer, missing = convert_missing_indexer(indexer)

            if missing:

                # reindex the axis to the new value
                # and set inplace
                if self.ndim == 1:
                    index = self.obj.index
                    new_index = index.insert(len(index), indexer)

                    # we have a coerced indexer, e.g. a float
                    # that matches in an Int64Index, so
                    # we will not create a duplicate index, rather
                    # index to that element
                    # e.g. 0.0 -> 0
                    # GH12246
                    if index.is_unique:
                        new_indexer = index.get_indexer([new_index[-1]])
                        if (new_indexer != -1).any():
                            return self._setitem_with_indexer(new_indexer,
                                                              value)

                    # this preserves dtype of the value
                    new_values = Series([value])._values
                    if len(self.obj._values):
                        try:
                            new_values = np.concatenate([self.obj._values,
                                                         new_values])
                        except TypeError:
                            as_obj = self.obj.astype(object)
                            new_values = np.concatenate([as_obj,
                                                         new_values])
                    self.obj._data = self.obj._constructor(
                        new_values, index=new_index, name=self.obj.name)._data
                    self.obj._maybe_update_cacher(clear=True)
                    return self.obj

                elif self.ndim == 2:

                    # no columns and scalar
                    if not len(self.obj.columns):
                        raise ValueError("cannot set a frame with no defined "
                                         "columns")

                    # append a Series
                    if isinstance(value, Series):

                        value = value.reindex(index=self.obj.columns,
                                              copy=True)
                        value.name = indexer

                    # a list-list
                    else:

                        # must have conforming columns
                        if is_list_like_indexer(value):
                            if len(value) != len(self.obj.columns):
                                raise ValueError("cannot set a row with "
                                                 "mismatched columns")

                        value = Series(value, index=self.obj.columns,
                                       name=indexer)

                    self.obj._data = self.obj.append(value)._data
                    self.obj._maybe_update_cacher(clear=True)
                    return self.obj

                # set using setitem (Panel and > dims)
                elif self.ndim >= 3:
                    return self.obj.__setitem__(indexer, value)

        # set
        item_labels = self.obj._get_axis(info_axis)

        # align and set the values
        if take_split_path:

            if not isinstance(indexer, tuple):
                indexer = self._tuplify(indexer)

            if isinstance(value, ABCSeries):
                value = self._align_series(indexer, value)

            info_idx = indexer[info_axis]
            if is_integer(info_idx):
                info_idx = [info_idx]
            labels = item_labels[info_idx]

            # if we have a partial multiindex, then need to adjust the plane
            # indexer here
            if (len(labels) == 1 and
                    isinstance(self.obj[labels[0]].axes[0], MultiIndex)):
                item = labels[0]
                obj = self.obj[item]
                index = obj.index
                idx = indexer[:info_axis][0]

                plane_indexer = tuple([idx]) + indexer[info_axis + 1:]
                lplane_indexer = length_of_indexer(plane_indexer[0], index)

                # require that we are setting the right number of values that
                # we are indexing
                if is_list_like_indexer(value) and np.iterable(
                        value) and lplane_indexer != len(value):

                    if len(obj[idx]) != len(value):
                        raise ValueError("cannot set using a multi-index "
                                         "selection indexer with a different "
                                         "length than the value")

                    # make sure we have an ndarray
                    value = getattr(value, 'values', value).ravel()

                    # we can directly set the series here
                    # as we select a slice indexer on the mi
                    idx = index._convert_slice_indexer(idx)
                    obj._consolidate_inplace()
                    obj = obj.copy()
                    obj._data = obj._data.setitem(indexer=tuple([idx]),
                                                  value=value)
                    self.obj[item] = obj
                    return

            # non-mi
            else:
                plane_indexer = indexer[:info_axis] + indexer[info_axis + 1:]
                if info_axis > 0:
                    plane_axis = self.obj.axes[:info_axis][0]
                    lplane_indexer = length_of_indexer(plane_indexer[0],
                                                       plane_axis)
                else:
                    lplane_indexer = 0

            def setter(item, v):
                s = self.obj[item]
                pi = plane_indexer[0] if lplane_indexer == 1 else plane_indexer

                # perform the equivalent of a setitem on the info axis
                # as we have a null slice or a slice with full bounds
                # which means essentially reassign to the columns of a
                # multi-dim object
                # GH6149 (null slice), GH10408 (full bounds)
                if (isinstance(pi, tuple) and
                        all(com.is_null_slice(idx) or
                            com.is_full_slice(idx, len(self.obj))
                            for idx in pi)):
                    s = v
                else:
                    # set the item, possibly having a dtype change
                    s._consolidate_inplace()
                    s = s.copy()
                    s._data = s._data.setitem(indexer=pi, value=v)
                    s._maybe_update_cacher(clear=True)

                # reset the sliced object if unique
                self.obj[item] = s

            def can_do_equal_len():
                """ return True if we have an equal len settable """
                if (not len(labels) == 1 or not np.iterable(value) or
                        is_scalar(plane_indexer[0])):
                    return False

                l = len(value)
                item = labels[0]
                index = self.obj[item].index

                # equal len list/ndarray
                if len(index) == l:
                    return True
                elif lplane_indexer == l:
                    return True

                return False

            # we need an iterable, with a ndim of at least 1
            # eg. don't pass through np.array(0)
            if is_list_like_indexer(value) and getattr(value, 'ndim', 1) > 0:

                # we have an equal len Frame
                if isinstance(value, ABCDataFrame) and value.ndim > 1:
                    sub_indexer = list(indexer)
                    multiindex_indexer = isinstance(labels, MultiIndex)

                    for item in labels:
                        if item in value:
                            sub_indexer[info_axis] = item
                            v = self._align_series(
                                tuple(sub_indexer), value[item],
                                multiindex_indexer)
                        else:
                            v = np.nan

                        setter(item, v)

                # we have an equal len ndarray/convertible to our labels
                elif np.array(value).ndim == 2:

                    # note that this coerces the dtype if we are mixed
                    # GH 7551
                    value = np.array(value, dtype=object)
                    if len(labels) != value.shape[1]:
                        raise ValueError('Must have equal len keys and value '
                                         'when setting with an ndarray')

                    for i, item in enumerate(labels):

                        # setting with a list, recoerces
                        setter(item, value[:, i].tolist())

                # we have an equal len list/ndarray
                elif can_do_equal_len():
                    setter(labels[0], value)

                # per label values
                else:

                    if len(labels) != len(value):
                        raise ValueError('Must have equal len keys and value '
                                         'when setting with an iterable')

                    for item, v in zip(labels, value):
                        setter(item, v)
            else:

                # scalar
                for item in labels:
                    setter(item, value)

        else:
            if isinstance(indexer, tuple):
                indexer = maybe_convert_ix(*indexer)

                # if we are setting on the info axis ONLY
                # set using those methods to avoid block-splitting
                # logic here
                if (len(indexer) > info_axis and
                        is_integer(indexer[info_axis]) and
                        all(com.is_null_slice(idx)
                            for i, idx in enumerate(indexer)
                            if i != info_axis) and
                        item_labels.is_unique):
                    self.obj[item_labels[indexer[info_axis]]] = value
                    return

            if isinstance(value, (ABCSeries, dict)):
                # TODO(EA): ExtensionBlock.setitem this causes issues with
                # setting for extensionarrays that store dicts. Need to decide
                # if it's worth supporting that.
                value = self._align_series(indexer, Series(value))

            elif isinstance(value, ABCDataFrame):
                value = self._align_frame(indexer, value)

            if isinstance(value, ABCPanel):
                value = self._align_panel(indexer, value)

            # check for chained assignment
            self.obj._check_is_chained_assignment_possible()

            # actually do the set
            self.obj._consolidate_inplace()
            self.obj._data = self.obj._data.setitem(indexer=indexer,
                                                    value=value)
            self.obj._maybe_update_cacher(clear=True)

    def _align_series(self, indexer, ser, multiindex_indexer=False):
        """
        Parameters
        ----------
        indexer : tuple, slice, scalar
            The indexer used to get the locations that will be set to
            `ser`

        ser : pd.Series
            The values to assign to the locations specified by `indexer`

        multiindex_indexer : boolean, optional
            Defaults to False. Should be set to True if `indexer` was from
            a `pd.MultiIndex`, to avoid unnecessary broadcasting.


        Returns:
        --------
        `np.array` of `ser` broadcast to the appropriate shape for assignment
        to the locations selected by `indexer`

        """
        if isinstance(indexer, (slice, np.ndarray, list, Index)):
            indexer = tuple([indexer])

        if isinstance(indexer, tuple):

            # flatten np.ndarray indexers
            def ravel(i):
                return i.ravel() if isinstance(i, np.ndarray) else i
            indexer = tuple(map(ravel, indexer))

            aligners = [not com.is_null_slice(idx) for idx in indexer]
            sum_aligners = sum(aligners)
            single_aligner = sum_aligners == 1
            is_frame = self.obj.ndim == 2
            is_panel = self.obj.ndim >= 3
            obj = self.obj

            # are we a single alignable value on a non-primary
            # dim (e.g. panel: 1,2, or frame: 0) ?
            # hence need to align to a single axis dimension
            # rather that find all valid dims

            # frame
            if is_frame:
                single_aligner = single_aligner and aligners[0]

            # panel
            elif is_panel:
                single_aligner = (single_aligner and
                                  (aligners[1] or aligners[2]))

            # we have a frame, with multiple indexers on both axes; and a
            # series, so need to broadcast (see GH5206)
            if (sum_aligners == self.ndim and
                    all(is_sequence(_) for _ in indexer)):
                ser = ser.reindex(obj.axes[0][indexer[0]], copy=True)._values

                # single indexer
                if len(indexer) > 1 and not multiindex_indexer:
                    l = len(indexer[1])
                    ser = np.tile(ser, l).reshape(l, -1).T

                return ser

            for i, idx in enumerate(indexer):
                ax = obj.axes[i]

                # multiple aligners (or null slices)
                if is_sequence(idx) or isinstance(idx, slice):
                    if single_aligner and com.is_null_slice(idx):
                        continue
                    new_ix = ax[idx]
                    if not is_list_like_indexer(new_ix):
                        new_ix = Index([new_ix])
                    else:
                        new_ix = Index(new_ix)
                    if ser.index.equals(new_ix) or not len(new_ix):
                        return ser._values.copy()

                    return ser.reindex(new_ix)._values

                # 2 dims
                elif single_aligner and is_frame:

                    # reindex along index
                    ax = self.obj.axes[1]
                    if ser.index.equals(ax) or not len(ax):
                        return ser._values.copy()
                    return ser.reindex(ax)._values

                # >2 dims
                elif single_aligner:

                    broadcast = []
                    for n, labels in enumerate(self.obj._get_plane_axes(i)):

                        # reindex along the matching dimensions
                        if len(labels & ser.index):
                            ser = ser.reindex(labels)
                        else:
                            broadcast.append((n, len(labels)))

                    # broadcast along other dims
                    ser = ser._values.copy()
                    for (axis, l) in broadcast:
                        shape = [-1] * (len(broadcast) + 1)
                        shape[axis] = l
                        ser = np.tile(ser, l).reshape(shape)

                    if self.obj.ndim == 3:
                        ser = ser.T

                    return ser

        elif is_scalar(indexer):
            ax = self.obj._get_axis(1)

            if ser.index.equals(ax):
                return ser._values.copy()

            return ser.reindex(ax)._values

        raise ValueError('Incompatible indexer with Series')

    def _align_frame(self, indexer, df):
        is_frame = self.obj.ndim == 2
        is_panel = self.obj.ndim >= 3

        if isinstance(indexer, tuple):

            aligners = [not com.is_null_slice(idx) for idx in indexer]
            sum_aligners = sum(aligners)
            # TODO: single_aligner is not used
            single_aligner = sum_aligners == 1  # noqa

            idx, cols = None, None
            sindexers = []
            for i, ix in enumerate(indexer):
                ax = self.obj.axes[i]
                if is_sequence(ix) or isinstance(ix, slice):
                    if isinstance(ix, np.ndarray):
                        ix = ix.ravel()
                    if idx is None:
                        idx = ax[ix]
                    elif cols is None:
                        cols = ax[ix]
                    else:
                        break
                else:
                    sindexers.append(i)

            # panel
            if is_panel:

                # need to conform to the convention
                # as we are not selecting on the items axis
                # and we have a single indexer
                # GH 7763
                if len(sindexers) == 1 and sindexers[0] != 0:
                    df = df.T

                if idx is None:
                    idx = df.index
                if cols is None:
                    cols = df.columns

            if idx is not None and cols is not None:

                if df.index.equals(idx) and df.columns.equals(cols):
                    val = df.copy()._values
                else:
                    val = df.reindex(idx, columns=cols)._values
                return val

        elif ((isinstance(indexer, slice) or is_list_like_indexer(indexer)) and
              is_frame):
            ax = self.obj.index[indexer]
            if df.index.equals(ax):
                val = df.copy()._values
            else:

                # we have a multi-index and are trying to align
                # with a particular, level GH3738
                if (isinstance(ax, MultiIndex) and
                        isinstance(df.index, MultiIndex) and
                        ax.nlevels != df.index.nlevels):
                    raise TypeError("cannot align on a multi-index with out "
                                    "specifying the join levels")

                val = df.reindex(index=ax)._values
            return val

        elif is_scalar(indexer) and is_panel:
            idx = self.obj.axes[1]
            cols = self.obj.axes[2]

            # by definition we are indexing on the 0th axis
            # a passed in dataframe which is actually a transpose
            # of what is needed
            if idx.equals(df.index) and cols.equals(df.columns):
                return df.copy()._values

            return df.reindex(idx, columns=cols)._values

        raise ValueError('Incompatible indexer with DataFrame')

    def _align_panel(self, indexer, df):
        # TODO: is_frame, is_panel are unused
        is_frame = self.obj.ndim == 2  # noqa
        is_panel = self.obj.ndim >= 3  # noqa
        raise NotImplementedError("cannot set using an indexer with a Panel "
                                  "yet!")

    def _getitem_tuple(self, tup):
        try:
            return self._getitem_lowerdim(tup)
        except IndexingError:
            pass

        # no multi-index, so validate all of the indexers
        self._has_valid_tuple(tup)

        # ugly hack for GH #836
        if self._multi_take_opportunity(tup):
            return self._multi_take(tup)

        # no shortcut needed
        retval = self.obj
        for i, key in enumerate(tup):
            if i >= self.obj.ndim:
                raise IndexingError('Too many indexers')

            if com.is_null_slice(key):
                continue

            retval = getattr(retval, self.name)._getitem_axis(key, axis=i)

        return retval

    def _multi_take_opportunity(self, tup):
        """
        Check whether there is the possibility to use ``_multi_take``.
        Currently the limit is that all axes being indexed must be indexed with
        list-likes.

        Parameters
        ----------
        tup : tuple
            Tuple of indexers, one per axis

        Returns
        -------
        boolean: Whether the current indexing can be passed through _multi_take
        """
        if not all(is_list_like_indexer(x) for x in tup):
            return False

        # just too complicated
        if any(com.is_bool_indexer(x) for x in tup):
            return False

        return True

    def _multi_take(self, tup):
        """
        Create the indexers for the passed tuple of keys, and execute the take
        operation. This allows the take operation to be executed all at once -
        rather than once for each dimension - improving efficiency.

        Parameters
        ----------
        tup : tuple
            Tuple of indexers, one per axis

        Returns
        -------
        values: same type as the object being indexed
        """
        # GH 836
        o = self.obj
        d = {axis: self._get_listlike_indexer(key, axis)
             for (key, axis) in zip(tup, o._AXIS_ORDERS)}
        return o._reindex_with_indexers(d, copy=True, allow_dups=True)

    def _convert_for_reindex(self, key, axis=None):
        return key

    def _handle_lowerdim_multi_index_axis0(self, tup):
        # we have an axis0 multi-index, handle or raise

        try:
            # fast path for series or for tup devoid of slices
            return self._get_label(tup, axis=self.axis)
        except TypeError:
            # slices are unhashable
            pass
        except Exception as e1:
            if isinstance(tup[0], (slice, Index)):
                raise IndexingError("Handle elsewhere")

            # raise the error if we are not sorted
            ax0 = self.obj._get_axis(0)
            if not ax0.is_lexsorted_for_tuple(tup):
                raise e1

        return None

    def _getitem_lowerdim(self, tup):

        # we can directly get the axis result since the axis is specified
        if self.axis is not None:
            axis = self.obj._get_axis_number(self.axis)
            return self._getitem_axis(tup, axis=axis)

        # we may have a nested tuples indexer here
        if self._is_nested_tuple_indexer(tup):
            return self._getitem_nested_tuple(tup)

        # we maybe be using a tuple to represent multiple dimensions here
        ax0 = self.obj._get_axis(0)
        # ...but iloc should handle the tuple as simple integer-location
        # instead of checking it as multiindex representation (GH 13797)
        if isinstance(ax0, MultiIndex) and self.name != 'iloc':
            result = self._handle_lowerdim_multi_index_axis0(tup)
            if result is not None:
                return result

        if len(tup) > self.obj.ndim:
            raise IndexingError("Too many indexers. handle elsewhere")

        # to avoid wasted computation
        # df.ix[d1:d2, 0] -> columns first (True)
        # df.ix[0, ['C', 'B', A']] -> rows first (False)
        for i, key in enumerate(tup):
            if is_label_like(key) or isinstance(key, tuple):
                section = self._getitem_axis(key, axis=i)

                # we have yielded a scalar ?
                if not is_list_like_indexer(section):
                    return section

                elif section.ndim == self.ndim:
                    # we're in the middle of slicing through a MultiIndex
                    # revise the key wrt to `section` by inserting an _NS
                    new_key = tup[:i] + (_NS,) + tup[i + 1:]

                else:
                    new_key = tup[:i] + tup[i + 1:]

                    # unfortunately need an odious kludge here because of
                    # DataFrame transposing convention
                    if (isinstance(section, ABCDataFrame) and i > 0 and
                            len(new_key) == 2):
                        a, b = new_key
                        new_key = b, a

                    if len(new_key) == 1:
                        new_key, = new_key

                # Slices should return views, but calling iloc/loc with a null
                # slice returns a new object.
                if com.is_null_slice(new_key):
                    return section
                # This is an elided recursive call to iloc/loc/etc'
                return getattr(section, self.name)[new_key]

        raise IndexingError('not applicable')

    def _getitem_nested_tuple(self, tup):
        # we have a nested tuple so have at least 1 multi-index level
        # we should be able to match up the dimensionaility here

        # we have too many indexers for our dim, but have at least 1
        # multi-index dimension, try to see if we have something like
        # a tuple passed to a series with a multi-index
        if len(tup) > self.ndim:
            result = self._handle_lowerdim_multi_index_axis0(tup)
            if result is not None:
                return result

            # this is a series with a multi-index specified a tuple of
            # selectors
            return self._getitem_axis(tup, axis=self.axis)

        # handle the multi-axis by taking sections and reducing
        # this is iterative
        obj = self.obj
        axis = 0
        for i, key in enumerate(tup):

            if com.is_null_slice(key):
                axis += 1
                continue

            current_ndim = obj.ndim
            obj = getattr(obj, self.name)._getitem_axis(key, axis=axis)
            axis += 1

            # if we have a scalar, we are done
            if is_scalar(obj) or not hasattr(obj, 'ndim'):
                break

            # has the dim of the obj changed?
            # GH 7199
            if obj.ndim < current_ndim:

                # GH 7516
                # if had a 3 dim and are going to a 2d
                # axes are reversed on a DataFrame
                if i >= 1 and current_ndim == 3 and obj.ndim == 2:
                    obj = obj.T

                axis -= 1

        return obj

    def _getitem_axis(self, key, axis=None):

        if axis is None:
            axis = self.axis or 0

        if is_iterator(key):
            key = list(key)
        self._validate_key(key, axis)

        labels = self.obj._get_axis(axis)
        if isinstance(key, slice):
            return self._get_slice_axis(key, axis=axis)
        elif (is_list_like_indexer(key) and
              not (isinstance(key, tuple) and
                   isinstance(labels, MultiIndex))):

            if hasattr(key, 'ndim') and key.ndim > 1:
                raise ValueError('Cannot index with multidimensional key')

            return self._getitem_iterable(key, axis=axis)
        else:

            # maybe coerce a float scalar to integer
            key = labels._maybe_cast_indexer(key)

            if is_integer(key):
                if axis == 0 and isinstance(labels, MultiIndex):
                    try:
                        return self._get_label(key, axis=axis)
                    except (KeyError, TypeError):
                        if self.obj.index.levels[0].is_integer():
                            raise

                # this is the fallback! (for a non-float, non-integer index)
                if not labels.is_floating() and not labels.is_integer():
                    return self._get_loc(key, axis=axis)

            return self._get_label(key, axis=axis)

    def _get_listlike_indexer(self, key, axis, raise_missing=False):
        """
        Transform a list-like of keys into a new index and an indexer.

        Parameters
        ----------
        key : list-like
            Target labels
        axis: int
            Dimension on which the indexing is being made
        raise_missing: bool
            Whether to raise a KeyError if some labels are not found. Will be
            removed in the future, and then this method will always behave as
            if raise_missing=True.

        Raises
        ------
        KeyError
            If at least one key was requested but none was found, and
            raise_missing=True.

        Returns
        -------
        keyarr: Index
            New index (coinciding with 'key' if the axis is unique)
        values : array-like
            An indexer for the return object; -1 denotes keys not found
        """
        o = self.obj
        ax = o._get_axis(axis)

        # Have the index compute an indexer or return None
        # if it cannot handle:
        indexer, keyarr = ax._convert_listlike_indexer(key,
                                                       kind=self.name)
        # We only act on all found values:
        if indexer is not None and (indexer != -1).all():
            self._validate_read_indexer(key, indexer, axis,
                                        raise_missing=raise_missing)
            return ax[indexer], indexer

        if ax.is_unique:
            # If we are trying to get actual keys from empty Series, we
            # patiently wait for a KeyError later on - otherwise, convert
            if len(ax) or not len(key):
                key = self._convert_for_reindex(key, axis)
            indexer = ax.get_indexer_for(key)
            keyarr = ax.reindex(keyarr)[0]
        else:
            keyarr, indexer, new_indexer = ax._reindex_non_unique(keyarr)

        self._validate_read_indexer(keyarr, indexer,
                                    o._get_axis_number(axis),
                                    raise_missing=raise_missing)
        return keyarr, indexer

    def _getitem_iterable(self, key, axis=None):
        """
        Index current object with an an iterable key (which can be a boolean
        indexer, or a collection of keys).

        Parameters
        ----------
        key : iterable
            Target labels, or boolean indexer
        axis: int, default None
            Dimension on which the indexing is being made

        Raises
        ------
        KeyError
            If no key was found. Will change in the future to raise if not all
            keys were found.
        IndexingError
            If the boolean indexer is unalignable with the object being
            indexed.

        Returns
        -------
        scalar, DataFrame, or Series: indexed value(s),
        """

        if axis is None:
            axis = self.axis or 0

        self._validate_key(key, axis)

        labels = self.obj._get_axis(axis)

        if com.is_bool_indexer(key):
            # A boolean indexer
            key = check_bool_indexer(labels, key)
            inds, = key.nonzero()
            return self.obj._take(inds, axis=axis)
        else:
            # A collection of keys
            keyarr, indexer = self._get_listlike_indexer(key, axis,
                                                         raise_missing=False)
            return self.obj._reindex_with_indexers({axis: [keyarr, indexer]},
                                                   copy=True, allow_dups=True)

    def _validate_read_indexer(self, key, indexer, axis, raise_missing=False):
        """
        Check that indexer can be used to return a result (e.g. at least one
        element was found, unless the list of keys was actually empty).

        Parameters
        ----------
        key : list-like
            Target labels (only used to show correct error message)
        indexer: array-like of booleans
            Indices corresponding to the key (with -1 indicating not found)
        axis: int
            Dimension on which the indexing is being made
        raise_missing: bool
            Whether to raise a KeyError if some labels are not found. Will be
            removed in the future, and then this method will always behave as
            if raise_missing=True.

        Raises
        ------
        KeyError
            If at least one key was requested but none was found, and
            raise_missing=True.
        """

        ax = self.obj._get_axis(axis)

        if len(key) == 0:
            return

        # Count missing values:
        missing = (indexer < 0).sum()

        if missing:
            if missing == len(indexer):
                raise KeyError(
                    u"None of [{key}] are in the [{axis}]".format(
                        key=key, axis=self.obj._get_axis_name(axis)))

            # We (temporarily) allow for some missing keys with .loc, except in
            # some cases (e.g. setting) in which "raise_missing" will be False
            if not(self.name == 'loc' and not raise_missing):
                not_found = list(set(key) - set(ax))
                raise KeyError("{} not in index".format(not_found))

            # we skip the warning on Categorical/Interval
            # as this check is actually done (check for
            # non-missing values), but a bit later in the
            # code, so we want to avoid warning & then
            # just raising

            _missing_key_warning = textwrap.dedent("""
            Passing list-likes to .loc or [] with any missing label will raise
            KeyError in the future, you can use .reindex() as an alternative.

            See the documentation here:
            https://pandas.pydata.org/pandas-docs/stable/indexing.html#deprecate-loc-reindex-listlike""")  # noqa

            if not (ax.is_categorical() or ax.is_interval()):
                warnings.warn(_missing_key_warning,
                              FutureWarning, stacklevel=6)

    def _convert_to_indexer(self, obj, axis=None, is_setter=False,
                            raise_missing=False):
        """
        Convert indexing key into something we can use to do actual fancy
        indexing on an ndarray

        Examples
        ix[:5] -> slice(0, 5)
        ix[[1,2,3]] -> [1,2,3]
        ix[['foo', 'bar', 'baz']] -> [i, j, k] (indices of foo, bar, baz)

        Going by Zen of Python?
        'In the face of ambiguity, refuse the temptation to guess.'
        raise AmbiguousIndexError with integer labels?
        - No, prefer label-based indexing
        """
        if axis is None:
            axis = self.axis or 0

        labels = self.obj._get_axis(axis)

        if isinstance(obj, slice):
            return self._convert_slice_indexer(obj, axis)

        # try to find out correct indexer, if not type correct raise
        try:
            obj = self._convert_scalar_indexer(obj, axis)
        except TypeError:

            # but we will allow setting
            if is_setter:
                pass

        # see if we are positional in nature
        is_int_index = labels.is_integer()
        is_int_positional = is_integer(obj) and not is_int_index

        # if we are a label return me
        try:
            return labels.get_loc(obj)
        except LookupError:
            if isinstance(obj, tuple) and isinstance(labels, MultiIndex):
                if is_setter and len(obj) == labels.nlevels:
                    return {'key': obj}
                raise
        except TypeError:
            pass
        except (ValueError):
            if not is_int_positional:
                raise

        # a positional
        if is_int_positional:

            # if we are setting and its not a valid location
            # its an insert which fails by definition
            if is_setter:

                # always valid
                if self.name == 'loc':
                    return {'key': obj}

                # a positional
                if (obj >= self.obj.shape[axis] and
                        not isinstance(labels, MultiIndex)):
                    raise ValueError("cannot set by positional indexing with "
                                     "enlargement")

            return obj

        if is_nested_tuple(obj, labels):
            return labels.get_locs(obj)

        elif is_list_like_indexer(obj):

            if com.is_bool_indexer(obj):
                obj = check_bool_indexer(labels, obj)
                inds, = obj.nonzero()
                return inds
            else:
                # When setting, missing keys are not allowed, even with .loc:
                kwargs = {'raise_missing': True if is_setter else
                          raise_missing}
                return self._get_listlike_indexer(obj, axis, **kwargs)[1]
        else:
            try:
                return labels.get_loc(obj)
            except LookupError:
                # allow a not found key only if we are a setter
                if not is_list_like_indexer(obj) and is_setter:
                    return {'key': obj}
                raise

    def _tuplify(self, loc):
        tup = [slice(None, None) for _ in range(self.ndim)]
        tup[0] = loc
        return tuple(tup)

    def _get_slice_axis(self, slice_obj, axis=None):
        obj = self.obj

        if axis is None:
            axis = self.axis or 0

        if not need_slice(slice_obj):
            return obj.copy(deep=False)
        indexer = self._convert_slice_indexer(slice_obj, axis)

        if isinstance(indexer, slice):
            return self._slice(indexer, axis=axis, kind='iloc')
        else:
            return self.obj._take(indexer, axis=axis)


class _IXIndexer(_NDFrameIndexer):
    """A primarily label-location based indexer, with integer position
    fallback.

    Warning: Starting in 0.20.0, the .ix indexer is deprecated, in
    favor of the more strict .iloc and .loc indexers.

    ``.ix[]`` supports mixed integer and label based access. It is
    primarily label based, but will fall back to integer positional
    access unless the corresponding axis is of integer type.

    ``.ix`` is the most general indexer and will support any of the
    inputs in ``.loc`` and ``.iloc``. ``.ix`` also supports floating
    point label schemes. ``.ix`` is exceptionally useful when dealing
    with mixed positional and label based hierarchical indexes.

    However, when an axis is integer based, ONLY label based access
    and not positional access is supported. Thus, in such cases, it's
    usually better to be explicit and use ``.iloc`` or ``.loc``.

    See more at :ref:`Advanced Indexing <advanced>`.

    """

    def __init__(self, name, obj):

        _ix_deprecation_warning = textwrap.dedent("""
            .ix is deprecated. Please use
            .loc for label based indexing or
            .iloc for positional indexing

            See the documentation here:
            http://pandas.pydata.org/pandas-docs/stable/indexing.html#ix-indexer-is-deprecated""")  # noqa

        warnings.warn(_ix_deprecation_warning,
                      DeprecationWarning, stacklevel=2)
        super(_IXIndexer, self).__init__(name, obj)

    @Appender(_NDFrameIndexer._validate_key.__doc__)
    def _validate_key(self, key, axis):
        if isinstance(key, slice):
            return True

        elif com.is_bool_indexer(key):
            return True

        elif is_list_like_indexer(key):
            return True

        else:

            self._convert_scalar_indexer(key, axis)

        return True

    def _convert_for_reindex(self, key, axis=None):
        """
        Transform a list of keys into a new array ready to be used as axis of
        the object we return (e.g. including NaNs).

        Parameters
        ----------
        key : list-like
            Target labels
        axis: int
            Where the indexing is being made

        Returns
        -------
        list-like of labels
        """

        if axis is None:
            axis = self.axis or 0
        labels = self.obj._get_axis(axis)

        if com.is_bool_indexer(key):
            key = check_bool_indexer(labels, key)
            return labels[key]

        if isinstance(key, Index):
            keyarr = labels._convert_index_indexer(key)
        else:
            # asarray can be unsafe, NumPy strings are weird
            keyarr = com._asarray_tuplesafe(key)

        if is_integer_dtype(keyarr):
            # Cast the indexer to uint64 if possible so
            # that the values returned from indexing are
            # also uint64.
            keyarr = labels._convert_arr_indexer(keyarr)

            if not labels.is_integer():
                keyarr = _ensure_platform_int(keyarr)
                return labels.take(keyarr)

        return keyarr


class _LocationIndexer(_NDFrameIndexer):
    _exception = Exception

    def __getitem__(self, key):
        if type(key) is tuple:
            key = tuple(com._apply_if_callable(x, self.obj)
                        for x in key)
            try:
                if self._is_scalar_access(key):
                    return self._getitem_scalar(key)
            except (KeyError, IndexError):
                pass
            return self._getitem_tuple(key)
        else:
            # we by definition only have the 0th axis
            axis = self.axis or 0

            maybe_callable = com._apply_if_callable(key, self.obj)
            return self._getitem_axis(maybe_callable, axis=axis)

    def _is_scalar_access(self, key):
        raise NotImplementedError()

    def _getitem_scalar(self, key):
        raise NotImplementedError()

    def _getitem_axis(self, key, axis=None):
        raise NotImplementedError()

    def _getbool_axis(self, key, axis=None):
        if axis is None:
            axis = self.axis or 0
        labels = self.obj._get_axis(axis)
        key = check_bool_indexer(labels, key)
        inds, = key.nonzero()
        try:
            return self.obj._take(inds, axis=axis)
        except Exception as detail:
            raise self._exception(detail)

    def _get_slice_axis(self, slice_obj, axis=None):
        """ this is pretty simple as we just have to deal with labels """
        if axis is None:
            axis = self.axis or 0

        obj = self.obj
        if not need_slice(slice_obj):
            return obj.copy(deep=False)

        labels = obj._get_axis(axis)
        indexer = labels.slice_indexer(slice_obj.start, slice_obj.stop,
                                       slice_obj.step, kind=self.name)

        if isinstance(indexer, slice):
            return self._slice(indexer, axis=axis, kind='iloc')
        else:
            return self.obj._take(indexer, axis=axis)


class _LocIndexer(_LocationIndexer):
    """
    Access a group of rows and columns by label(s) or a boolean array.

    ``.loc[]`` is primarily label based, but may also be used with a
    boolean array.

    Allowed inputs are:

    - A single label, e.g. ``5`` or ``'a'``, (note that ``5`` is
      interpreted as a *label* of the index, and **never** as an
      integer position along the index).
    - A list or array of labels, e.g. ``['a', 'b', 'c']``.
    - A slice object with labels, e.g. ``'a':'f'``.

      .. warning:: Note that contrary to usual python slices, **both** the
          start and the stop are included

    - A boolean array of the same length as the axis being sliced,
      e.g. ``[True, False, True]``.
    - A ``callable`` function with one argument (the calling Series, DataFrame
      or Panel) and that returns valid output for indexing (one of the above)

    See more at :ref:`Selection by Label <indexing.label>`

    See Also
    --------
    DataFrame.at : Access a single value for a row/column label pair
    DataFrame.iloc : Access group of rows and columns by integer position(s)
    DataFrame.xs : Returns a cross-section (row(s) or column(s)) from the
        Series/DataFrame.
    Series.loc : Access group of values using labels

    Examples
    --------
    **Getting values**

    >>> df = pd.DataFrame([[1, 2], [4, 5], [7, 8]],
    ...      index=['cobra', 'viper', 'sidewinder'],
    ...      columns=['max_speed', 'shield'])
    >>> df
                max_speed  shield
    cobra               1       2
    viper               4       5
    sidewinder          7       8

    Single label. Note this returns the row as a Series.

    >>> df.loc['viper']
    max_speed    4
    shield       5
    Name: viper, dtype: int64

    List of labels. Note using ``[[]]`` returns a DataFrame.

    >>> df.loc[['viper', 'sidewinder']]
                max_speed  shield
    viper               4       5
    sidewinder          7       8

    Single label for row and column

    >>> df.loc['cobra', 'shield']
    2

    Slice with labels for row and single label for column. As mentioned
    above, note that both the start and stop of the slice are included.

    >>> df.loc['cobra':'viper', 'max_speed']
    cobra    1
    viper    4
    Name: max_speed, dtype: int64

    Boolean list with the same length as the row axis

    >>> df.loc[[False, False, True]]
                max_speed  shield
    sidewinder          7       8

    Conditional that returns a boolean Series

    >>> df.loc[df['shield'] > 6]
                max_speed  shield
    sidewinder          7       8

    Conditional that returns a boolean Series with column labels specified

    >>> df.loc[df['shield'] > 6, ['max_speed']]
                max_speed
    sidewinder          7

    Callable that returns a boolean Series

    >>> df.loc[lambda df: df['shield'] == 8]
                max_speed  shield
    sidewinder          7       8

    **Setting values**

    Set value for all items matching the list of labels

    >>> df.loc[['viper', 'sidewinder'], ['shield']] = 50
    >>> df
                max_speed  shield
    cobra               1       2
    viper               4      50
    sidewinder          7      50

    Set value for an entire row

    >>> df.loc['cobra'] = 10
    >>> df
                max_speed  shield
    cobra              10      10
    viper               4      50
    sidewinder          7      50

    Set value for an entire column

    >>> df.loc[:, 'max_speed'] = 30
    >>> df
                max_speed  shield
    cobra              30      10
    viper              30      50
    sidewinder         30      50

    Set value for rows matching callable condition

    >>> df.loc[df['shield'] > 35] = 0
    >>> df
                max_speed  shield
    cobra              30      10
    viper               0       0
    sidewinder          0       0

    **Getting values on a DataFrame with an index that has integer labels**

    Another example using integers for the index

    >>> df = pd.DataFrame([[1, 2], [4, 5], [7, 8]],
    ...      index=[7, 8, 9], columns=['max_speed', 'shield'])
    >>> df
       max_speed  shield
    7          1       2
    8          4       5
    9          7       8

    Slice with integer labels for rows. As mentioned above, note that both
    the start and stop of the slice are included.

    >>> df.loc[7:9]
       max_speed  shield
    7          1       2
    8          4       5
    9          7       8

    **Getting values with a MultiIndex**

    A number of examples using a DataFrame with a MultiIndex

    >>> tuples = [
    ...    ('cobra', 'mark i'), ('cobra', 'mark ii'),
    ...    ('sidewinder', 'mark i'), ('sidewinder', 'mark ii'),
    ...    ('viper', 'mark ii'), ('viper', 'mark iii')
    ... ]
    >>> index = pd.MultiIndex.from_tuples(tuples)
    >>> values = [[12, 2], [0, 4], [10, 20],
    ...         [1, 4], [7, 1], [16, 36]]
    >>> df = pd.DataFrame(values, columns=['max_speed', 'shield'], index=index)
    >>> df
                         max_speed  shield
    cobra      mark i           12       2
               mark ii           0       4
    sidewinder mark i           10      20
               mark ii           1       4
    viper      mark ii           7       1
               mark iii         16      36

    Single label. Note this returns a DataFrame with a single index.

    >>> df.loc['cobra']
             max_speed  shield
    mark i          12       2
    mark ii          0       4

    Single index tuple. Note this returns a Series.

    >>> df.loc[('cobra', 'mark ii')]
    max_speed    0
    shield       4
    Name: (cobra, mark ii), dtype: int64

    Single label for row and column. Similar to passing in a tuple, this
    returns a Series.

    >>> df.loc['cobra', 'mark i']
    max_speed    12
    shield        2
    Name: (cobra, mark i), dtype: int64

    Single tuple. Note using ``[[]]`` returns a DataFrame.

    >>> df.loc[[('cobra', 'mark ii')]]
                   max_speed  shield
    cobra mark ii          0       4

    Single tuple for the index with a single label for the column

    >>> df.loc[('cobra', 'mark i'), 'shield']
    2

    Slice from index tuple to single label

    >>> df.loc[('cobra', 'mark i'):'viper']
                         max_speed  shield
    cobra      mark i           12       2
               mark ii           0       4
    sidewinder mark i           10      20
               mark ii           1       4
    viper      mark ii           7       1
               mark iii         16      36

    Slice from index tuple to index tuple

    >>> df.loc[('cobra', 'mark i'):('viper', 'mark ii')]
                        max_speed  shield
    cobra      mark i          12       2
               mark ii          0       4
    sidewinder mark i          10      20
               mark ii          1       4
    viper      mark ii          7       1

    Raises
    ------
    KeyError:
        when any items are not found
    """

    _valid_types = ("labels (MUST BE IN THE INDEX), slices of labels (BOTH "
                    "endpoints included! Can be slices of integers if the "
                    "index is integers), listlike of labels, boolean")
    _exception = KeyError

    @Appender(_NDFrameIndexer._validate_key.__doc__)
    def _validate_key(self, key, axis):

        # valid for a collection of labels (we check their presence later)
        # slice of labels (where start-end in labels)
        # slice of integers (only if in the labels)
        # boolean

        if isinstance(key, slice):
            return

        if com.is_bool_indexer(key):
            return

        if not is_list_like_indexer(key):
            self._convert_scalar_indexer(key, axis)

    def _is_scalar_access(self, key):
        # this is a shortcut accessor to both .loc and .iloc
        # that provide the equivalent access of .at and .iat
        # a) avoid getting things via sections and (to minimize dtype changes)
        # b) provide a performant path
        if not hasattr(key, '__len__'):
            return False

        if len(key) != self.ndim:
            return False

        for i, k in enumerate(key):
            if not is_scalar(k):
                return False

            ax = self.obj.axes[i]
            if isinstance(ax, MultiIndex):
                return False

            if not ax.is_unique:
                return False

        return True

    def _getitem_scalar(self, key):
        # a fast-path to scalar access
        # if not, raise
        values = self.obj._get_value(*key)
        return values

    def _get_partial_string_timestamp_match_key(self, key, labels):
        """Translate any partial string timestamp matches in key, returning the
        new key (GH 10331)"""
        if isinstance(labels, MultiIndex):
            if isinstance(key, compat.string_types) and \
                    labels.levels[0].is_all_dates:
                # Convert key '2016-01-01' to
                # ('2016-01-01'[, slice(None, None, None)]+)
                key = tuple([key] + [slice(None)] * (len(labels.levels) - 1))

            if isinstance(key, tuple):
                # Convert (..., '2016-01-01', ...) in tuple to
                # (..., slice('2016-01-01', '2016-01-01', None), ...)
                new_key = []
                for i, component in enumerate(key):
                    if isinstance(component, compat.string_types) and \
                            labels.levels[i].is_all_dates:
                        new_key.append(slice(component, component, None))
                    else:
                        new_key.append(component)
                key = tuple(new_key)

        return key

    def _getitem_axis(self, key, axis=None):
        if axis is None:
            axis = self.axis or 0

        if is_iterator(key):
            key = list(key)

        labels = self.obj._get_axis(axis)
        key = self._get_partial_string_timestamp_match_key(key, labels)

        if isinstance(key, slice):
            self._validate_key(key, axis)
            return self._get_slice_axis(key, axis=axis)
        elif com.is_bool_indexer(key):
            return self._getbool_axis(key, axis=axis)
        elif is_list_like_indexer(key):

            # convert various list-like indexers
            # to a list of keys
            # we will use the *values* of the object
            # and NOT the index if its a PandasObject
            if isinstance(labels, MultiIndex):

                if isinstance(key, (ABCSeries, np.ndarray)) and key.ndim <= 1:
                    # Series, or 0,1 ndim ndarray
                    # GH 14730
                    key = list(key)
                elif isinstance(key, ABCDataFrame):
                    # GH 15438
                    raise NotImplementedError("Indexing a MultiIndex with a "
                                              "DataFrame key is not "
                                              "implemented")
                elif hasattr(key, 'ndim') and key.ndim > 1:
                    raise NotImplementedError("Indexing a MultiIndex with a "
                                              "multidimensional key is not "
                                              "implemented")

                if (not isinstance(key, tuple) and len(key) > 1 and
                        not isinstance(key[0], tuple)):
                    key = tuple([key])

            # an iterable multi-selection
            if not (isinstance(key, tuple) and isinstance(labels, MultiIndex)):

                if hasattr(key, 'ndim') and key.ndim > 1:
                    raise ValueError('Cannot index with multidimensional key')

                return self._getitem_iterable(key, axis=axis)

            # nested tuple slicing
            if is_nested_tuple(key, labels):
                locs = labels.get_locs(key)
                indexer = [slice(None)] * self.ndim
                indexer[axis] = locs
                return self.obj.iloc[tuple(indexer)]

        # fall thru to straight lookup
        self._validate_key(key, axis)
        return self._get_label(key, axis=axis)


class _iLocIndexer(_LocationIndexer):
    """
    Purely integer-location based indexing for selection by position.

    ``.iloc[]`` is primarily integer position based (from ``0`` to
    ``length-1`` of the axis), but may also be used with a boolean
    array.

    Allowed inputs are:

    - An integer, e.g. ``5``.
    - A list or array of integers, e.g. ``[4, 3, 0]``.
    - A slice object with ints, e.g. ``1:7``.
    - A boolean array.
    - A ``callable`` function with one argument (the calling Series, DataFrame
      or Panel) and that returns valid output for indexing (one of the above).
      This is useful in method chains, when you don't have a reference to the
      calling object, but would like to base your selection on some value.

    ``.iloc`` will raise ``IndexError`` if a requested indexer is
    out-of-bounds, except *slice* indexers which allow out-of-bounds
    indexing (this conforms with python/numpy *slice* semantics).

    See more at ref:`Selection by Position <indexing.integer>`.

    See Also
    --------
    DataFrame.iat : Fast integer location scalar accessor.
    DataFrame.loc : Purely label-location based indexer for selection by label.
    Series.iloc : Purely integer-location based indexing for
                   selection by position.

    Examples
    --------

    >>> mydict = [{'a': 1, 'b': 2, 'c': 3, 'd': 4},
    ...           {'a': 100, 'b': 200, 'c': 300, 'd': 400},
    ...           {'a': 1000, 'b': 2000, 'c': 3000, 'd': 4000 }]
    >>> df = pd.DataFrame(mydict)
    >>> df
          a     b     c     d
    0     1     2     3     4
    1   100   200   300   400
    2  1000  2000  3000  4000

    **Indexing just the rows**

    With a scalar integer.

    >>> type(df.iloc[0])
    <class 'pandas.core.series.Series'>
    >>> df.iloc[0]
    a    1
    b    2
    c    3
    d    4
    Name: 0, dtype: int64

    With a list of integers.

    >>> df.iloc[[0]]
       a  b  c  d
    0  1  2  3  4
    >>> type(df.iloc[[0]])
    <class 'pandas.core.frame.DataFrame'>

    >>> df.iloc[[0, 1]]
         a    b    c    d
    0    1    2    3    4
    1  100  200  300  400

    With a `slice` object.

    >>> df.iloc[:3]
          a     b     c     d
    0     1     2     3     4
    1   100   200   300   400
    2  1000  2000  3000  4000

    With a boolean mask the same length as the index.

    >>> df.iloc[[True, False, True]]
          a     b     c     d
    0     1     2     3     4
    2  1000  2000  3000  4000

    With a callable, useful in method chains. The `x` passed
    to the ``lambda`` is the DataFrame being sliced. This selects
    the rows whose index label even.

    >>> df.iloc[lambda x: x.index % 2 == 0]
          a     b     c     d
    0     1     2     3     4
    2  1000  2000  3000  4000

    **Indexing both axes**

    You can mix the indexer types for the index and columns. Use ``:`` to
    select the entire axis.

    With scalar integers.

    >>> df.iloc[0, 1]
    2

    With lists of integers.

    >>> df.iloc[[0, 2], [1, 3]]
          b     d
    0     2     4
    2  2000  4000

    With `slice` objects.

    >>> df.iloc[1:3, 0:3]
          a     b     c
    1   100   200   300
    2  1000  2000  3000

    With a boolean array whose length matches the columns.

    >>> df.iloc[:, [True, False, True, False]]
          a     c
    0     1     3
    1   100   300
    2  1000  3000

    With a callable function that expects the Series or DataFrame.

    >>> df.iloc[:, lambda df: [0, 2]]
          a     c
    0     1     3
    1   100   300
    2  1000  3000
    """

    _valid_types = ("integer, integer slice (START point is INCLUDED, END "
                    "point is EXCLUDED), listlike of integers, boolean array")
    _exception = IndexError

    def _validate_key(self, key, axis):
        if com.is_bool_indexer(key):
            if hasattr(key, 'index') and isinstance(key.index, Index):
                if key.index.inferred_type == 'integer':
                    raise NotImplementedError("iLocation based boolean "
                                              "indexing on an integer type "
                                              "is not available")
                raise ValueError("iLocation based boolean indexing cannot use "
                                 "an indexable as a mask")
            return

        if isinstance(key, slice):
            return
        elif is_integer(key):
            self._validate_integer(key, axis)
        elif isinstance(key, tuple):
            # a tuple should already have been caught by this point
            # so don't treat a tuple as a valid indexer
            raise IndexingError('Too many indexers')
        elif is_list_like_indexer(key):
            # check that the key does not exceed the maximum size of the index
            arr = np.array(key)
            l = len(self.obj._get_axis(axis))

            if len(arr) and (arr.max() >= l or arr.min() < -l):
                raise IndexError("positional indexers are out-of-bounds")
        else:
            raise ValueError("Can only index by location with "
                             "a [{types}]".format(types=self._valid_types))

    def _has_valid_setitem_indexer(self, indexer):
        self._has_valid_positional_setitem_indexer(indexer)

    def _is_scalar_access(self, key):
        # this is a shortcut accessor to both .loc and .iloc
        # that provide the equivalent access of .at and .iat
        # a) avoid getting things via sections and (to minimize dtype changes)
        # b) provide a performant path
        if not hasattr(key, '__len__'):
            return False

        if len(key) != self.ndim:
            return False

        for i, k in enumerate(key):
            if not is_integer(k):
                return False

            ax = self.obj.axes[i]
            if not ax.is_unique:
                return False

        return True

    def _getitem_scalar(self, key):
        # a fast-path to scalar access
        # if not, raise
        values = self.obj._get_value(*key, takeable=True)
        return values

    def _validate_integer(self, key, axis):
        # return a boolean if we have a valid integer indexer

        ax = self.obj._get_axis(axis)
        l = len(ax)
        if key >= l or key < -l:
            raise IndexError("single positional indexer is out-of-bounds")

    def _getitem_tuple(self, tup):

        self._has_valid_tuple(tup)
        try:
            return self._getitem_lowerdim(tup)
        except:
            pass

        retval = self.obj
        axis = 0
        for i, key in enumerate(tup):
            if i >= self.obj.ndim:
                raise IndexingError('Too many indexers')

            if com.is_null_slice(key):
                axis += 1
                continue

            retval = getattr(retval, self.name)._getitem_axis(key, axis=axis)

            # if the dim was reduced, then pass a lower-dim the next time
            if retval.ndim < self.ndim:
                axis -= 1

            # try to get for the next axis
            axis += 1

        return retval

    def _get_slice_axis(self, slice_obj, axis=None):
        if axis is None:
            axis = self.axis or 0
        obj = self.obj

        if not need_slice(slice_obj):
            return obj.copy(deep=False)

        slice_obj = self._convert_slice_indexer(slice_obj, axis)
        if isinstance(slice_obj, slice):
            return self._slice(slice_obj, axis=axis, kind='iloc')
        else:
            return self.obj._take(slice_obj, axis=axis)

    def _get_list_axis(self, key, axis=None):
        """
        Return Series values by list or array of integers

        Parameters
        ----------
        key : list-like positional indexer
        axis : int (can only be zero)

        Returns
        -------
        Series object
        """
        if axis is None:
            axis = self.axis or 0
        try:
            return self.obj._take(key, axis=axis)
        except IndexError:
            # re-raise with different error message
            raise IndexError("positional indexers are out-of-bounds")

    def _getitem_axis(self, key, axis=None):
        if axis is None:
            axis = self.axis or 0

        if isinstance(key, slice):
            return self._get_slice_axis(key, axis=axis)

        if isinstance(key, list):
            key = np.asarray(key)

        if com.is_bool_indexer(key):
            self._validate_key(key, axis)
            return self._getbool_axis(key, axis=axis)

        # a list of integers
        elif is_list_like_indexer(key):
            return self._get_list_axis(key, axis=axis)

        # a single integer
        else:
            key = self._convert_scalar_indexer(key, axis)

            if not is_integer(key):
                raise TypeError("Cannot index by location index with a "
                                "non-integer key")

            # validate the location
            self._validate_integer(key, axis)

            return self._get_loc(key, axis=axis)

    def _convert_to_indexer(self, obj, axis=None, is_setter=False):
        """ much simpler as we only have to deal with our valid types """
        if axis is None:
            axis = self.axis or 0

        # make need to convert a float key
        if isinstance(obj, slice):
            return self._convert_slice_indexer(obj, axis)

        elif is_float(obj):
            return self._convert_scalar_indexer(obj, axis)

        try:
            self._validate_key(obj, axis)
            return obj
        except ValueError:
            raise ValueError("Can only index by location with "
                             "a [{types}]".format(types=self._valid_types))


class _ScalarAccessIndexer(_NDFrameIndexer):
    """ access scalars quickly """

    def _convert_key(self, key, is_setter=False):
        return list(key)

    def __getitem__(self, key):
        if not isinstance(key, tuple):

            # we could have a convertible item here (e.g. Timestamp)
            if not is_list_like_indexer(key):
                key = tuple([key])
            else:
                raise ValueError('Invalid call for scalar access (getting)!')

        key = self._convert_key(key)
        return self.obj._get_value(*key, takeable=self._takeable)

    def __setitem__(self, key, value):
        if isinstance(key, tuple):
            key = tuple(com._apply_if_callable(x, self.obj)
                        for x in key)
        else:
            # scalar callable may return tuple
            key = com._apply_if_callable(key, self.obj)

        if not isinstance(key, tuple):
            key = self._tuplify(key)
        if len(key) != self.obj.ndim:
            raise ValueError('Not enough indexers for scalar access '
                             '(setting)!')
        key = list(self._convert_key(key, is_setter=True))
        key.append(value)
        self.obj._set_value(*key, takeable=self._takeable)


class _AtIndexer(_ScalarAccessIndexer):
    """
    Access a single value for a row/column label pair.

    Similar to ``loc``, in that both provide label-based lookups. Use
    ``at`` if you only need to get or set a single value in a DataFrame
    or Series.

    See Also
    --------
    DataFrame.iat : Access a single value for a row/column pair by integer
        position
    DataFrame.loc : Access a group of rows and columns by label(s)
    Series.at : Access a single value using a label

    Examples
    --------
    >>> df = pd.DataFrame([[0, 2, 3], [0, 4, 1], [10, 20, 30]],
    ...                   index=[4, 5, 6], columns=['A', 'B', 'C'])
    >>> df
        A   B   C
    4   0   2   3
    5   0   4   1
    6  10  20  30

    Get value at specified row/column pair

    >>> df.at[4, 'B']
    2

    Set value at specified row/column pair

    >>> df.at[4, 'B'] = 10
    >>> df.at[4, 'B']
    10

    Get value within a Series

    >>> df.loc[5].at['B']
    4

    Raises
    ------
    KeyError
        When label does not exist in DataFrame
    """

    _takeable = False

    def _convert_key(self, key, is_setter=False):
        """ require they keys to be the same type as the index (so we don't
        fallback)
        """

        # allow arbitrary setting
        if is_setter:
            return list(key)

        for ax, i in zip(self.obj.axes, key):
            if ax.is_integer():
                if not is_integer(i):
                    raise ValueError("At based indexing on an integer index "
                                     "can only have integer indexers")
            else:
                if is_integer(i):
                    raise ValueError("At based indexing on an non-integer "
                                     "index can only have non-integer "
                                     "indexers")
        return key


class _iAtIndexer(_ScalarAccessIndexer):
    """
    Access a single value for a row/column pair by integer position.

    Similar to ``iloc``, in that both provide integer-based lookups. Use
    ``iat`` if you only need to get or set a single value in a DataFrame
    or Series.

    See Also
    --------
    DataFrame.at : Access a single value for a row/column label pair
    DataFrame.loc : Access a group of rows and columns by label(s)
    DataFrame.iloc : Access a group of rows and columns by integer position(s)

    Examples
    --------
    >>> df = pd.DataFrame([[0, 2, 3], [0, 4, 1], [10, 20, 30]],
    ...                   columns=['A', 'B', 'C'])
    >>> df
        A   B   C
    0   0   2   3
    1   0   4   1
    2  10  20  30

    Get value at specified row/column pair

    >>> df.iat[1, 2]
    1

    Set value at specified row/column pair

    >>> df.iat[1, 2] = 10
    >>> df.iat[1, 2]
    10

    Get value within a series

    >>> df.loc[0].iat[1]
    2

    Raises
    ------
    IndexError
        When integer position is out of bounds
    """

    _takeable = True

    def _has_valid_setitem_indexer(self, indexer):
        self._has_valid_positional_setitem_indexer(indexer)

    def _convert_key(self, key, is_setter=False):
        """ require integer args (and convert to label arguments) """
        for a, i in zip(self.obj.axes, key):
            if not is_integer(i):
                raise ValueError("iAt based indexing can only have integer "
                                 "indexers")
        return key


def length_of_indexer(indexer, target=None):
    """return the length of a single non-tuple indexer which could be a slice
    """
    if target is not None and isinstance(indexer, slice):
        l = len(target)
        start = indexer.start
        stop = indexer.stop
        step = indexer.step
        if start is None:
            start = 0
        elif start < 0:
            start += l
        if stop is None or stop > l:
            stop = l
        elif stop < 0:
            stop += l
        if step is None:
            step = 1
        elif step < 0:
            step = -step
        return (stop - start + step - 1) // step
    elif isinstance(indexer, (ABCSeries, Index, np.ndarray, list)):
        return len(indexer)
    elif not is_list_like_indexer(indexer):
        return 1
    raise AssertionError("cannot find the length of the indexer")


def convert_to_index_sliceable(obj, key):
    """if we are index sliceable, then return my slicer, otherwise return None
    """
    idx = obj.index
    if isinstance(key, slice):
        return idx._convert_slice_indexer(key, kind='getitem')

    elif isinstance(key, compat.string_types):

        # we are an actual column
        if obj._data.items.contains(key):
            return None

        # We might have a datetimelike string that we can translate to a
        # slice here via partial string indexing
        if idx.is_all_dates:
            try:
                return idx._get_string_slice(key)
            except (KeyError, ValueError, NotImplementedError):
                return None

    return None


def check_bool_indexer(ax, key):
    # boolean indexing, need to check that the data are aligned, otherwise
    # disallowed

    # this function assumes that is_bool_indexer(key) == True

    result = key
    if isinstance(key, ABCSeries) and not key.index.equals(ax):
        result = result.reindex(ax)
        mask = isna(result._values)
        if mask.any():
            raise IndexingError('Unalignable boolean Series provided as '
                                'indexer (index of the boolean Series and of '
                                'the indexed object do not match')
        result = result.astype(bool)._values
    elif is_sparse(result):
        result = result.to_dense()
        result = np.asarray(result, dtype=bool)
    else:
        # is_bool_indexer has already checked for nulls in the case of an
        # object array key, so no check needed here
        result = np.asarray(result, dtype=bool)

    return result


def check_setitem_lengths(indexer, value, values):
    """Validate that value and indexer are the same length.

    An special-case is allowed for when the indexer is a boolean array
    and the number of true values equals the length of ``value``. In
    this case, no exception is raised.

    Parameters
    ----------
    indexer : sequence
        The key for the setitem
    value : array-like
        The value for the setitem
    values : array-like
        The values being set into

    Returns
    -------
    None

    Raises
    ------
    ValueError
        When the indexer is an ndarray or list and the lengths don't
        match.
    """
    # boolean with truth values == len of the value is ok too
    if isinstance(indexer, (np.ndarray, list)):
        if is_list_like(value) and len(indexer) != len(value):
            if not (isinstance(indexer, np.ndarray) and
                    indexer.dtype == np.bool_ and
                    len(indexer[indexer]) == len(value)):
                raise ValueError("cannot set using a list-like indexer "
                                 "with a different length than the value")
    # slice
    elif isinstance(indexer, slice):

        if is_list_like(value) and len(values):
            if len(value) != length_of_indexer(indexer, values):
                raise ValueError("cannot set using a slice indexer with a "
                                 "different length than the value")


def convert_missing_indexer(indexer):
    """ reverse convert a missing indexer, which is a dict
    return the scalar indexer and a boolean indicating if we converted
    """

    if isinstance(indexer, dict):

        # a missing key (but not a tuple indexer)
        indexer = indexer['key']

        if isinstance(indexer, bool):
            raise KeyError("cannot use a single bool to index into setitem")
        return indexer, True

    return indexer, False


def convert_from_missing_indexer_tuple(indexer, axes):
    """ create a filtered indexer that doesn't have any missing indexers """

    def get_indexer(_i, _idx):
        return (axes[_i].get_loc(_idx['key']) if isinstance(_idx, dict) else
                _idx)

    return tuple(get_indexer(_i, _idx) for _i, _idx in enumerate(indexer))


def maybe_convert_indices(indices, n):
    """
    Attempt to convert indices into valid, positive indices.

    If we have negative indices, translate to positive here.
    If we have indices that are out-of-bounds, raise an IndexError.

    Parameters
    ----------
    indices : array-like
        The array of indices that we are to convert.
    n : int
        The number of elements in the array that we are indexing.

    Returns
    -------
    valid_indices : array-like
        An array-like of positive indices that correspond to the ones
        that were passed in initially to this function.

    Raises
    ------
    IndexError : one of the converted indices either exceeded the number
        of elements (specified by `n`) OR was still negative.
    """
    indices = deepcopy(indices)

    if isinstance(indices, list):
        indices = np.array(indices)
        if len(indices) == 0:
            # If list is empty, np.array will return float and cause indexing
            # errors.
            return np.empty(0, dtype=np.intp)

    mask = indices < 0
    if mask.any():
        # indices is np.array, so we don't use deepcopy
        indices = np.copy(indices)
        indices[mask] += n

    mask = (indices >= n) | (indices < 0)
    if mask.any():
        raise IndexError("indices are out-of-bounds")
    return indices


def validate_indices(indices, n):
    """Perform bounds-checking for an indexer.

    -1 is allowed for indicating missing values.

    Parameters
    ----------
    indices : ndarray
    n : int
        length of the array being indexed

    Raises
    ------
    ValueError

    Examples
    --------
    >>> validate_indices([1, 2], 3)
    # OK
    >>> validate_indices([1, -2], 3)
    ValueError
    >>> validate_indices([1, 2, 3], 3)
    IndexError
    >>> validate_indices([-1, -1], 0)
    # OK
    >>> validate_indices([0, 1], 0)
    IndexError
    """
    if len(indices):
        min_idx = indices.min()
        if min_idx < -1:
            msg = ("'indices' contains values less than allowed ({} < {})"
                   .format(min_idx, -1))
            raise ValueError(msg)

        max_idx = indices.max()
        if max_idx >= n:
            raise IndexError("indices are out-of-bounds")


def maybe_convert_ix(*args):
    """
    We likely want to take the cross-product
    """

    ixify = True
    for arg in args:
        if not isinstance(arg, (np.ndarray, list, ABCSeries, Index)):
            ixify = False

    if ixify:
        return np.ix_(*args)
    else:
        return args


def is_nested_tuple(tup, labels):
    # check for a compatible nested tuple and multiindexes among the axes
    if not isinstance(tup, tuple):
        return False

    for i, k in enumerate(tup):

        if is_list_like(k) or isinstance(k, slice):
            return isinstance(labels, MultiIndex)

    return False


def is_list_like_indexer(key):
    # allow a list_like, but exclude NamedTuples which can be indexers
    return is_list_like(key) and not (isinstance(key, tuple) and
                                      type(key) is not tuple)


def is_label_like(key):
    # select a label or row
    return not isinstance(key, slice) and not is_list_like_indexer(key)


def need_slice(obj):
    return (obj.start is not None or obj.stop is not None or
            (obj.step is not None and obj.step != 1))


def maybe_droplevels(index, key):
    # drop levels
    original_index = index
    if isinstance(key, tuple):
        for _ in key:
            try:
                index = index.droplevel(0)
            except:
                # we have dropped too much, so back out
                return original_index
    else:
        try:
            index = index.droplevel(0)
        except:
            pass

    return index


def _non_reducing_slice(slice_):
    """
    Ensurse that a slice doesn't reduce to a Series or Scalar.

    Any user-paseed `subset` should have this called on it
    to make sure we're always working with DataFrames.
    """
    # default to column slice, like DataFrame
    # ['A', 'B'] -> IndexSlices[:, ['A', 'B']]
    kinds = tuple(list(compat.string_types) + [ABCSeries, np.ndarray, Index,
                                               list])
    if isinstance(slice_, kinds):
        slice_ = IndexSlice[:, slice_]

    def pred(part):
        # true when slice does *not* reduce
        return isinstance(part, slice) or is_list_like(part)

    if not is_list_like(slice_):
        if not isinstance(slice_, slice):
            # a 1-d slice, like df.loc[1]
            slice_ = [[slice_]]
        else:
            # slice(a, b, c)
            slice_ = [slice_]  # to tuplize later
    else:
        slice_ = [part if pred(part) else [part] for part in slice_]
    return tuple(slice_)


def _maybe_numeric_slice(df, slice_, include_bool=False):
    """
    want nice defaults for background_gradient that don't break
    with non-numeric data. But if slice_ is passed go with that.
    """
    if slice_ is None:
        dtypes = [np.number]
        if include_bool:
            dtypes.append(bool)
        slice_ = IndexSlice[:, df.select_dtypes(include=dtypes).columns]
    return slice_