move pandas.api.lib.infer_dtype -> pandas.api.types

move union_categoricals to pandas.api.types

make feather single test

fix odd tz hashing issue

remove circular import

pandas.core.common warning message

Author: jreback

doc/source/categorical.rst

@@ -673,7 +673,7 @@ will be the union of the categories being combined.
 
 .. ipython:: python
 
-    from pandas.types.concat import union_categoricals
+    from pandas.api.types import union_categoricals
     a = pd.Categorical(["b", "c"])
     b = pd.Categorical(["a", "b"])
     union_categoricals([a, b])

doc/source/whatsnew/v0.20.0.txt

@@ -489,7 +489,7 @@ Other Enhancements
 
 
 - ``TimedeltaIndex`` now has a custom datetick formatter specifically designed for nanosecond level precision (:issue:`8711`)
-- ``pd.types.concat.union_categoricals`` gained the ``ignore_ordered`` argument to allow ignoring the ordered attribute of unioned categoricals (:issue:`13410`). See the :ref:`categorical union docs <categorical.union>` for more information.
+- ``pd.api.types.union_categoricals`` gained the ``ignore_ordered`` argument to allow ignoring the ordered attribute of unioned categoricals (:issue:`13410`). See the :ref:`categorical union docs <categorical.union>` for more information.
 - ``DataFrame.to_latex()`` and ``DataFrame.to_string()`` now allow optional header aliases. (:issue:`15536`)
 - Re-enable the ``parse_dates`` keyword of ``pd.read_excel()`` to parse string columns as dates (:issue:`14326`)
 - Added ``.empty`` property to subclasses of ``Index``. (:issue:`15270`)
@@ -1057,6 +1057,8 @@ If indicated, a deprecation warning will be issued if you reference that module.
     "pandas._hash", "pandas.tools.libhash", ""
     "pandas._window", "pandas.core.libwindow", ""
 
+- The function :func:`~pandas.api.type.union_categoricals` is now importable from ``pandas.api.types``, formerly from ``pandas.types.concat`` (:issue:`15998`)
+
 .. _whatsnew_0200.api_breaking.sort_index:
 
 DataFrame.sort_index changes

pandas/api/types/__init__.py

@@ -1,4 +1,5 @@
 """ public toolkit API """
 
 from pandas.core.typed.api import *  # noqa
+from pandas.core.typed.concat import union_categoricals  # noqa
 del np  # noqa

pandas/core/algorithms.py

@@ -28,7 +28,7 @@
 from pandas.compat.numpy import _np_version_under1p10
 from pandas.core.typed.missing import isnull
 
-import pandas.core.common as com
+from pandas.core import common as com
 from pandas.compat import string_types
 from pandas._libs import algos, lib, hashtable as htable
 from pandas._libs.tslib import iNaT

pandas/core/common.py

@@ -60,7 +60,7 @@ def wrapper(*args, **kwargs):
             warnings.warn("pandas.core.common.{t} is deprecated. "
                           "These are not longer public API functions, "
                           "but can be imported from "
-                          "pandas.core.typed.common.{t} instead".format(t=t),
+                          "pandas.api.types.{t} instead".format(t=t),
                           DeprecationWarning, stacklevel=3)
             return getattr(common, t)(*args, **kwargs)
         return wrapper

pandas/core/series.py

@@ -2869,8 +2869,6 @@ def _sanitize_index(data, index, copy=False):
         data = data.asobject
     elif isinstance(data, DatetimeIndex):
         data = data._to_embed(keep_tz=True)
-        if copy:
-            data = data.copy()
     elif isinstance(data, np.ndarray):
 
         # coerce datetimelike types

pandas/core/typed/concat.py

@@ -5,18 +5,18 @@
 import numpy as np
 import pandas._libs.tslib as tslib
 from pandas import compat
-from pandas.core.algorithms import take_1d
-from .common import (is_categorical_dtype,
-                     is_sparse,
-                     is_datetimetz,
-                     is_datetime64_dtype,
-                     is_timedelta64_dtype,
-                     is_period_dtype,
-                     is_object_dtype,
-                     is_bool_dtype,
-                     is_dtype_equal,
-                     _NS_DTYPE,
-                     _TD_DTYPE)
+from pandas.core.typed.common import (
+    is_categorical_dtype,
+    is_sparse,
+    is_datetimetz,
+    is_datetime64_dtype,
+    is_timedelta64_dtype,
+    is_period_dtype,
+    is_object_dtype,
+    is_bool_dtype,
+    is_dtype_equal,
+    _NS_DTYPE,
+    _TD_DTYPE)
 from pandas.core.typed.generic import (
     ABCDatetimeIndex, ABCTimedeltaIndex,
     ABCPeriodIndex)
@@ -277,6 +277,8 @@ def _maybe_unwrap(x):
         if sort_categories and not categories.is_monotonic_increasing:
             categories = categories.sort_values()
             indexer = categories.get_indexer(first.categories)
+
+            from pandas.core.algorithms import take_1d
             new_codes = take_1d(indexer, new_codes, fill_value=-1)
     elif ignore_order or all(not c.ordered for c in to_union):
         # different categories - union and recode
@@ -289,6 +291,8 @@ def _maybe_unwrap(x):
         for c in to_union:
             if len(c.categories) > 0:
                 indexer = categories.get_indexer(c.categories)
+
+                from pandas.core.algorithms import take_1d
                 new_codes.append(take_1d(indexer, c.codes, fill_value=-1))
             else:
                 # must be all NaN

pandas/core/typed/dtypes.py

@@ -209,8 +209,15 @@ def __new__(cls, unit=None, tz=None):
             raise ValueError("DatetimeTZDtype constructor must have a tz "
                              "supplied")
 
+        # hash with the actual tz if we can
+        # some cannot be hashed, so stringfy
+        try:
+            key = (unit, tz)
+            hash(key)
+        except TypeError:
+            key = (unit, str(tz))
+
         # set/retrieve from cache
-        key = (unit, str(tz))
         try:
             return cls._cache[key]
         except KeyError:

pandas/tests/api/test_lib.py

@@ -1,7 +1,9 @@
 # -*- coding: utf-8 -*-
 
+from warnings import catch_warnings
 import numpy as np
 
+import pandas
 from pandas.core import common as com
 from pandas.api import types
 from pandas.util import testing as tm
@@ -28,7 +30,7 @@ class TestTypes(Base, tm.TestCase):
                'is_dict_like', 'is_iterator', 'is_file_like',
                'is_list_like', 'is_hashable',
                'is_named_tuple', 'is_sequence',
-               'pandas_dtype']
+               'pandas_dtype', 'union_categoricals']
 
     def test_types(self):
 
@@ -82,3 +84,10 @@ def test_removed_from_core_common(self):
         for t in ['is_null_datelike_scalar',
                   'ensure_float']:
             self.assertRaises(AttributeError, lambda: getattr(com, t))
+
+
+def test_moved_infer_dtype():
+
+    with catch_warnings(record=True):
+        e = pandas.lib.infer_dtype('foo')
+        assert e is not None

pandas/tests/api/test_types.py

@@ -12,21 +12,36 @@
 
 class TestDatetimeIndex(tm.TestCase):
 
+    def test_construction_caching(self):
+
+        df = pd.DataFrame({'dt': pd.date_range('20130101', periods=3),
+                           'dttz': pd.date_range('20130101', periods=3,
+                                                 tz='US/Eastern'),
+                           'dt_with_null': [pd.Timestamp('20130101'), pd.NaT,
+                                            pd.Timestamp('20130103')],
+                           'dtns': pd.date_range('20130101', periods=3,
+                                                 freq='ns')})
+        assert df.dttz.dtype.tz.zone == 'US/Eastern'
+
     def test_construction_with_alt(self):
 
         i = pd.date_range('20130101', periods=5, freq='H', tz='US/Eastern')
         i2 = DatetimeIndex(i, dtype=i.dtype)
         self.assert_index_equal(i, i2)
+        assert i.tz.zone == 'US/Eastern'
 
         i2 = DatetimeIndex(i.tz_localize(None).asi8, tz=i.dtype.tz)
         self.assert_index_equal(i, i2)
+        assert i.tz.zone == 'US/Eastern'
 
         i2 = DatetimeIndex(i.tz_localize(None).asi8, dtype=i.dtype)
         self.assert_index_equal(i, i2)
+        assert i.tz.zone == 'US/Eastern'
 
         i2 = DatetimeIndex(
             i.tz_localize(None).asi8, dtype=i.dtype, tz=i.dtype.tz)
         self.assert_index_equal(i, i2)
+        assert i.tz.zone == 'US/Eastern'
 
         # localize into the provided tz
         i2 = DatetimeIndex(i.tz_localize(None).asi8, tz='UTC')

pandas/tests/indexes/datetimes/test_construction.py

@@ -11,6 +11,7 @@
 from pandas.util.testing import assert_frame_equal, ensure_clean
 
 
+@pytest.mark.single
 class TestFeather(object):
 
     def check_error_on_write(self, df, exc):
@@ -52,6 +53,7 @@ def test_basic(self):
                            'dtns': pd.date_range('20130101', periods=3,
                                                  freq='ns')})
 
+        assert df.dttz.dtype.tz.zone == 'US/Eastern'
         self.check_round_trip(df)
 
     def test_strided_data_issues(self):

pandas/tests/io/test_feather.py

@@ -636,7 +636,6 @@ def pxd(name):
       packages=['pandas',
                 'pandas.api',
                 'pandas.api.types',
-                'pandas.api.lib',
                 'pandas.compat',
                 'pandas.compat.numpy',
                 'pandas.core',