BUG: Fixes problem with SparseArray coercing to float if index is passed

pandas/core/frame.py

@@ -2649,7 +2649,7 @@ def insert(self, loc, column, value, allow_duplicates=False):
                           allow_duplicates=allow_duplicates)
 
     def assign(self, **kwargs):
-        r"""
+        """
         Assign new columns to a DataFrame, returning a new object
         (a copy) with all the original columns in addition to the new ones.
 
@@ -2714,8 +2714,7 @@ def assign(self, **kwargs):
         9  10 -0.758542  2.302585
         """
 
-        # See GH19163
-        data = self.copy().to_dense()
+        data = self.copy()
 
         # do all calculations first...
         results = OrderedDict()

pandas/core/sparse/array.py

@@ -27,7 +27,7 @@
     is_scalar, is_dtype_equal)
 from pandas.core.dtypes.cast import (
     maybe_convert_platform, maybe_promote,
-    astype_nansafe, find_common_type)
+    astype_nansafe, find_common_type, infer_dtype_from)
 from pandas.core.dtypes.missing import isna, notna, na_value_for_dtype
 
 import pandas._libs.sparse as splib
@@ -195,7 +195,7 @@ def __new__(cls, data, sparse_index=None, index=None, kind='integer',
                 data = np.nan
             if not is_scalar(data):
                 raise Exception("must only pass scalars with an index ")
-            values = np.empty(len(index), dtype='float64')
+            values = np.empty(len(index), dtype=infer_dtype_from(data)[0])
             values.fill(data)
             data = values
 

pandas/tests/frame/test_mutate_columns.py

@@ -6,9 +6,9 @@
 import numpy as np
 from pandas.compat import PY36
 
-from pandas import DataFrame, Series, Index, MultiIndex
+from pandas import DataFrame, Series, Index, MultiIndex, SparseSeries
 
-from pandas.util.testing import assert_frame_equal
+from pandas.util.testing import assert_frame_equal, assert_series_equal
 
 import pandas.util.testing as tm
 
@@ -55,13 +55,6 @@ def test_assign(self):
         result = df.assign(A=lambda x: x.A + x.B)
         assert_frame_equal(result, expected)
 
-        # SparseDataFrame
-        # See GH 19163
-        result = df.to_sparse(fill_value=False).assign(newcol=False)
-        expected = df.assign(newcol=False)
-        assert type(result) is DataFrame
-        assert_frame_equal(expected, result)
-
     def test_assign_multiple(self):
         df = DataFrame([[1, 4], [2, 5], [3, 6]], columns=['A', 'B'])
         result = df.assign(C=[7, 8, 9], D=df.A, E=lambda x: x.B)

pandas/tests/sparse/frame/test_frame.py

@@ -1271,3 +1271,15 @@ def test_quantile_multi(self):
 
         tm.assert_frame_equal(result, dense_expected)
         tm.assert_sp_frame_equal(result, sparse_expected)
+
+    def test_assign_with_sparse_frame(self):
+        # GH 19163
+        df = pd.DataFrame({"a":[1,2,3]})
+        res = df.to_sparse(fill_value=False).assign(newcol=False)
+        exp = df.assign(newcol=False).to_sparse(fill_value=False)
+
+        tm.assert_sp_frame_equal(res, exp)
+
+        for column in res.columns:
+            assert type(res[column]) is SparseSeries
+

pandas/tests/sparse/test_array.py

@@ -113,6 +113,17 @@ def test_constructor_spindex_dtype(self):
         assert arr.dtype == np.int64
         assert arr.fill_value == 0
 
+    @pytest.mark.parametrize('scalar,dtype', [(False, bool), (0.0, 'float64'), (1, 'int64'), ('z', 'object')])
+    def test_scalar_with_index_infer_dtype(self, scalar, dtype):
+        # GH 19163
+        arr = SparseArray(scalar, index=[1,2,3], fill_value=scalar)
+        exp = SparseArray([scalar, scalar, scalar], fill_value=scalar)
+
+        tm.assert_sp_array_equal(arr, exp)
+
+        assert arr.dtype == dtype
+        assert exp.dtype == dtype
+
     def test_sparseseries_roundtrip(self):
         # GH 13999
         for kind in ['integer', 'block']: