rather insidious issue with taking from a uint64 array with missing values :>

were being promoted to floats

Author: jreback

.gitignore

@@ -19,6 +19,7 @@
 .noseids
 .ipynb_checkpoints
 .tags
+.cache/
 
 # Compiled source #
 ###################

pandas/tests/frame/test_reshape.py

@@ -629,7 +629,6 @@ def verify(df):
  names=[None, 'B'])
 
  right = DataFrame(vals, columns=cols, index=idx)
- import pdb; pdb.set_trace()
  assert_frame_equal(left, right)
 
  # GH4862

pandas/tools/hashing.py

@@ -137,7 +137,7 @@ def hash_tuples(vals, encoding='utf8', hash_key=None):
  is_tuple = True
  elif not is_list_like(vals):
  raise TypeError("must be convertible to a list-of-tuples")
- import pdb; pdb.set_trace()
+
  if not isinstance(vals, MultiIndex):
  vals = MultiIndex.from_tuples(vals)
 

pandas/tools/tests/test_hashing.py

@@ -152,6 +152,17 @@ def test_categorical_consistency(self):
  tm.assert_series_equal(h1, h2)
  tm.assert_series_equal(h1, h3)
 
+ def test_categorical_with_nan_consistency(self):
+ c = pd.Categorical.from_codes(
+ [-1, 0, 1, 2, 3, 4],
+ categories=pd.date_range('2012-01-01', periods=5, name='B'))
+ expected = hash_array(c, categorize=False)
+ c = pd.Categorical.from_codes(
+ [0],
+ categories=[pd.Timestamp('2012-01-01')])
+ result = hash_array(c, categorize=False)
+ assert result.item() in expected
+
  def test_pandas_errors(self):
 
  for obj in [pd.Timestamp('20130101'), tm.makePanel()]:

pandas/types/cast.py

@@ -12,7 +12,8 @@
  is_datetime64tz_dtype, is_datetime64_dtype,
  is_timedelta64_dtype, is_dtype_equal,
  is_float_dtype, is_complex_dtype,
- is_integer_dtype, is_datetime_or_timedelta_dtype,
+ is_integer_dtype, is_unsigned_integer_dtype,
+ is_datetime_or_timedelta_dtype,
  is_bool_dtype, is_scalar,
  _string_dtypes,
  _coerce_to_dtype,
@@ -265,6 +266,9 @@ def _maybe_promote(dtype, fill_value=np.nan):
  elif is_datetimetz(dtype):
  if isnull(fill_value):
  fill_value = iNaT
+ elif is_unsigned_integer_dtype(dtype):
+ if isnull(fill_value):
+ fill_value = np.iinfo(np.uint64).max
  elif is_float(fill_value):
  if issubclass(dtype.type, np.bool_):
  dtype = np.object_