API: allow scalar setting/getting via float indexer on integer indexes

closes #12333

Author: Jeff Reback <jeff@reback.net>

Closes #12370 from jreback/float and squashes the following commits:

3a148a7 [Jeff Reback] API: allow scalar setting/getting via float indexer on integer indexes

Author: jreback

doc/source/advanced.rst

@@ -717,6 +717,10 @@ values NOT in the categories, similarly to how you can reindex ANY pandas index.
 Int64Index and RangeIndex
 ~~~~~~~~~~~~~~~~~~~~~~~~~
 
+.. warning::
+
+   Indexing on an integer-based Index with floats has been clarified in 0.18.0, for a summary of the changes, see :ref:`here <whatsnew_0180.float_indexers>`.
+
 ``Int64Index`` is a fundamental basic index in *pandas*. This is an Immutable array implementing an ordered, sliceable set.
 Prior to 0.18.0, the ``Int64Index`` would provide the default index for all ``NDFrame`` objects.
 
@@ -736,7 +740,6 @@ Float64Index
    operations by about 30x and boolean indexing operations on the
    ``Float64Index`` itself are about 2x as fast.
 
-
 .. versionadded:: 0.13.0
 
 By default a ``Float64Index`` will be automatically created when passing floating, or mixed-integer-floating values in index creation.
@@ -797,12 +800,12 @@ In non-float indexes, slicing using floats will raise a ``TypeError``
 
 .. warning::
 
-   Using a scalar float indexer has been removed in 0.18.0, so the following will raise a ``TypeError``
+   Using a scalar float indexer for ``.iloc`` has been removed in 0.18.0, so the following will raise a ``TypeError``
 
    .. code-block:: python
 
-      In [3]: pd.Series(range(5))[3.0]
-      TypeError: cannot do label indexing on <class 'pandas.indexes.range.RangeIndex'> with these indexers [3.0] of <type 'float'>
+      In [3]: pd.Series(range(5)).iloc[3.0]
+      TypeError: cannot do positional indexing on <class 'pandas.indexes.range.RangeIndex'> with these indexers [3.0] of <type 'float'>
 
    Further the treatment of ``.ix`` with a float indexer on a non-float index, will be label based, and thus coerce the index.
 

doc/source/indexing.rst

@@ -53,6 +53,10 @@ advanced indexing.
    but instead subclass ``PandasObject``, similarly to the rest of the pandas objects. This should be
    a transparent change with only very limited API implications (See the :ref:`Internal Refactoring <whatsnew_0150.refactoring>`)
 
+.. warning::
+
+   Indexing on an integer-based Index with floats has been clarified in 0.18.0, for a summary of the changes, see :ref:`here <whatsnew_0180.float_indexers>`.
+
 See the :ref:`MultiIndex / Advanced Indexing <advanced>` for ``MultiIndex`` and more advanced indexing documentation.
 
 See the :ref:`cookbook<cookbook.selection>` for some advanced strategies

doc/source/whatsnew/v0.18.0.txt

@@ -1024,11 +1024,11 @@ Removal of deprecated float indexers
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 In :issue:`4892` indexing with floating point numbers on a non-``Float64Index`` was deprecated (in version 0.14.0).
-In 0.18.0, this deprecation warning is removed and these will now raise a ``TypeError``. (:issue:`12165`)
+In 0.18.0, this deprecation warning is removed and these will now raise a ``TypeError``. (:issue:`12165`, :issue:`12333`)
 
 .. ipython:: python
 
-   s = pd.Series([1,2,3])
+   s = pd.Series([1, 2, 3], index=[4, 5, 6])
    s
    s2 = pd.Series([1, 2, 3], index=list('abc'))
    s2
@@ -1037,15 +1037,18 @@ Previous Behavior:
 
 .. code-block:: python
 
-   In [2]: s[1.0]
+   # this is label indexing
+   In [2]: s[5.0]
    FutureWarning: scalar indexers for index type Int64Index should be integers and not floating point
    Out[2]: 2
 
+   # this is positional indexing
    In [3]: s.iloc[1.0]
    FutureWarning: scalar indexers for index type Int64Index should be integers and not floating point
    Out[3]: 2
 
-   In [4]: s.loc[1.0]
+   # this is label indexing
+   In [4]: s.loc[5.0]
    FutureWarning: scalar indexers for index type Int64Index should be integers and not floating point
    Out[4]: 2
 
@@ -1062,33 +1065,54 @@ Previous Behavior:
 
 New Behavior:
 
+For iloc, getting & setting via a float scalar will always raise.
+
 .. code-block:: python
 
-   In [2]: s[1.0]
-   TypeError: cannot do label indexing on <class 'pandas.indexes.range.RangeIndex'> with these indexers [1.0] of <type 'float'>
+   In [3]: s.iloc[2.0]
+   TypeError: cannot do label indexing on <class 'pandas.indexes.numeric.Int64Index'> with these indexers [2.0] of <type 'float'>
 
-   In [3]: s.iloc[1.0]
-   TypeError: cannot do label indexing on <class 'pandas.indexes.range.RangeIndex'> with these indexers [1.0] of <type 'float'>
+Other indexers will coerce to a like integer for both getting and setting. The ``FutureWarning`` has been dropped for ``.loc``, ``.ix`` and ``[]``.
 
-   In [4]: s.loc[1.0]
-   TypeError: cannot do label indexing on <class 'pandas.indexes.range.RangeIndex'> with these indexers [1.0] of <type 'float'>
+.. ipython:: python
 
-   # .ix will now cause this to be a label lookup and coerce to and Index
-   In [5]: s2.ix[1.0] = 10
+   s[5.0]
+   s.loc[5.0]
+   s.ix[5.0]
 
-   In [6]: s2
-   Out[3]:
-   a       1
-   b       2
-   c       3
-   1.0    10
-   dtype: int64
+and setting
+
+.. ipython:: python
+
+   s_copy = s.copy()
+   s_copy[5.0] = 10
+   s_copy
+   s_copy = s.copy()
+   s_copy.loc[5.0] = 10
+   s_copy
+   s_copy = s.copy()
+   s_copy.ix[5.0] = 10
+   s_copy
+
+Slicing will also coerce integer-like floats to integers for a non-``Float64Index``.
+
+.. ipython:: python
+
+   s.loc[5.0:6]
+   s.ix[5.0:6]
+
+Note that for floats that are NOT coercible to ints, the label based bounds will be excluded
+
+.. ipython:: python
+
+   s.loc[5.1:6]
+   s.ix[5.1:6]
 
 Float indexing on a ``Float64Index`` is unchanged.
 
 .. ipython:: python
 
-   s = pd.Series([1,2,3],index=np.arange(3.))
+   s = pd.Series([1, 2, 3], index=np.arange(3.))
    s[1.0]
    s[1.0:2.5]
 

pandas/core/frame.py

@@ -2016,7 +2016,7 @@ def _getitem_array(self, key):
             # with all other indexing behavior
             if isinstance(key, Series) and not key.index.equals(self.index):
                 warnings.warn("Boolean Series key will be reindexed to match "
-                              "DataFrame index.", UserWarning)
+                              "DataFrame index.", UserWarning, stacklevel=3)
             elif len(key) != len(self.index):
                 raise ValueError('Item wrong length %d instead of %d.' %
                                  (len(key), len(self.index)))

pandas/core/indexing.py

@@ -995,6 +995,10 @@ def _getitem_axis(self, key, axis=0):
 
             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: