ENH: Allow setitem in a frame where only mixed numerics are present

RELEASE.rst

@@ -62,6 +62,7 @@ pandas 0.11.0
     strings that can be parsed with datetime.strptime
   - Add ``axes`` property to ``Series`` for compatibility 
   - Add ``xs`` function to ``Series`` for compatibility 
+  - Allow setitem in a frame where only mixed numerics are present (e.g. int and float), (GH3037_)
 
 **API Changes**
 
@@ -182,6 +183,7 @@ pandas 0.11.0
 .. _GH3010: https://github.com/pydata/pandas/issues/3010
 .. _GH3012: https://github.com/pydata/pandas/issues/3012
 .. _GH3029: https://github.com/pydata/pandas/issues/3029
+.. _GH3037: https://github.com/pydata/pandas/issues/3037
 .. _GH3041: https://github.com/pydata/pandas/issues/3041
 
 

pandas/core/frame.py

@@ -2121,7 +2121,8 @@ def _setitem_frame(self, key, value):
             raise ValueError('Must pass DataFrame with boolean values only')
 
         if self._is_mixed_type:
-            raise ValueError('Cannot do boolean setting on mixed-type frame')
+            if not self._is_numeric_mixed_type:
+                raise ValueError('Cannot do boolean setting on mixed-type frame')
 
         self.where(-key, value, inplace=True)
 

pandas/core/generic.py

@@ -671,8 +671,11 @@ def consolidate(self, inplace=False):
 
     @property
     def _is_mixed_type(self):
-        self._consolidate_inplace()
-        return len(self._data.blocks) > 1
+        return self._data.is_mixed_type
+
+    @property
+    def _is_numeric_mixed_type(self):
+        return self._data.is_numeric_mixed_type
 
     def _reindex_axis(self, new_index, fill_method, axis, copy):
         new_data = self._data.reindex_axis(new_index, axis=axis,

pandas/core/internals.py

@@ -980,6 +980,16 @@ def _consolidate_check(self):
         self._is_consolidated = len(dtypes) == len(set(dtypes))
         self._known_consolidated = True
 
+    @property
+    def is_mixed_type(self):
+        self._consolidate_inplace()
+        return len(self.blocks) > 1
+
+    @property
+    def is_numeric_mixed_type(self):
+        self._consolidate_inplace()
+        return all([ block.is_numeric for block in self.blocks ])
+
     def get_numeric_data(self, copy=False, type_list=None, as_blocks = False):
         """
         Parameters
@@ -1227,9 +1237,10 @@ def consolidate(self):
         return BlockManager(new_blocks, self.axes)
 
     def _consolidate_inplace(self):
-        self.blocks = _consolidate(self.blocks, self.items)
-        self._is_consolidated = True
-        self._known_consolidated = True
+        if not self.is_consolidated():
+            self.blocks = _consolidate(self.blocks, self.items)
+            self._is_consolidated = True
+            self._known_consolidated = True
 
     def get(self, item):
         _, block = self._find_block(item)

pandas/core/reshape.py

@@ -159,7 +159,7 @@ def get_new_values(self):
             dtype, fill_value = _maybe_promote(values.dtype)
             new_values = np.empty(result_shape, dtype=dtype)
             new_values.fill(fill_value)
-
+ 
         new_mask = np.zeros(result_shape, dtype=bool)
 
         # is there a simpler / faster way of doing this?

pandas/tests/test_frame.py

@@ -8660,6 +8660,33 @@ def test_boolean_indexing(self):
         df1[df1 > 2.0 * df2] = -1
         assert_frame_equal(df1, expected)
 
+    def test_boolean_indexing_mixed(self):
+        df = DataFrame(
+            {0L: {35: np.nan, 40: np.nan, 43: np.nan, 49: np.nan, 50: np.nan},
+             1L: {35: np.nan,
+                  40: 0.32632316859446198,
+                  43: np.nan,
+                  49: 0.32632316859446198,
+                  50: 0.39114724480578139},
+             2L: {35: np.nan, 40: np.nan, 43: 0.29012581014105987, 49: np.nan, 50: np.nan},
+             3L: {35: np.nan, 40: np.nan, 43: np.nan, 49: np.nan, 50: np.nan},
+             4L: {35: 0.34215328467153283, 40: np.nan, 43: np.nan, 49: np.nan, 50: np.nan},
+             'y': {35: 0, 40: 0, 43: 0, 49: 0, 50: 1}})
+
+        # mixed int/float ok
+        df2 = df.copy()
+        df2[df2>0.3] = 1
+        expected = df.copy()
+        expected.loc[40,1] = 1
+        expected.loc[49,1] = 1
+        expected.loc[50,1] = 1
+        expected.loc[35,4] = 1
+        assert_frame_equal(df2,expected)
+
+        # add object, should this raise?
+        df['foo'] = 'test'
+        self.assertRaises(ValueError, df.__setitem__, df>0.3, 1)
+
     def test_sum_bools(self):
         df = DataFrame(index=range(1), columns=range(10))
         bools = isnull(df)