Mix EA into DTA/TDA; part of 24024 (pandas-dev#24502)

Author: jbrockmendel

pandas/core/arrays/datetimelike.py

@@ -28,13 +28,14 @@
 from pandas.core.dtypes.missing import isna
 
 from pandas.core import nanops
-from pandas.core.algorithms import checked_add_with_arr, take, unique1d
+from pandas.core.algorithms import (
+    checked_add_with_arr, take, unique1d, value_counts)
 import pandas.core.common as com
 
 from pandas.tseries import frequencies
 from pandas.tseries.offsets import DateOffset, Tick
 
-from .base import ExtensionOpsMixin
+from .base import ExtensionArray, ExtensionOpsMixin
 
 
 def _make_comparison_op(cls, op):
@@ -343,7 +344,9 @@ def ceil(self, freq, ambiguous='raise', nonexistent='raise'):
         return self._round(freq, RoundTo.PLUS_INFTY, ambiguous, nonexistent)
 
 
-class DatetimeLikeArrayMixin(ExtensionOpsMixin, AttributesMixin):
+class DatetimeLikeArrayMixin(ExtensionOpsMixin,
+                             AttributesMixin,
+                             ExtensionArray):
     """
     Shared Base/Mixin class for DatetimeArray, TimedeltaArray, PeriodArray
 
@@ -701,6 +704,43 @@ def repeat(self, repeats, *args, **kwargs):
         values = self._data.repeat(repeats)
         return type(self)(values.view('i8'), dtype=self.dtype)
 
+    def value_counts(self, dropna=False):
+        """
+        Return a Series containing counts of unique values.
+
+        Parameters
+        ----------
+        dropna : boolean, default True
+            Don't include counts of NaT values.
+
+        Returns
+        -------
+        Series
+        """
+        from pandas import Series, Index
+
+        if dropna:
+            values = self[~self.isna()]._data
+        else:
+            values = self._data
+
+        cls = type(self)
+
+        result = value_counts(values, sort=False, dropna=dropna)
+        index = Index(cls(result.index.view('i8'), dtype=self.dtype),
+                      name=result.index.name)
+        return Series(result.values, index=index, name=result.name)
+
+    def map(self, mapper):
+        # TODO(GH-23179): Add ExtensionArray.map
+        # Need to figure out if we want ExtensionArray.map first.
+        # If so, then we can refactor IndexOpsMixin._map_values to
+        # a standalone function and call from here..
+        # Else, just rewrite _map_infer_values to do the right thing.
+        from pandas import Index
+
+        return Index(self).map(mapper).array
+
     # ------------------------------------------------------------------
     # Null Handling
 
@@ -1357,10 +1397,9 @@ def _reduce(self, name, axis=0, skipna=True, **kwargs):
         if op:
             return op(axis=axis, skipna=skipna, **kwargs)
         else:
-            raise TypeError("cannot perform {name} with type {dtype}"
-                            .format(name=name, dtype=self.dtype))
-            # TODO: use super(DatetimeLikeArrayMixin, self)._reduce
-            #  after we subclass ExtensionArray
+            return super(DatetimeLikeArrayMixin, self)._reduce(
+                name, skipna, **kwargs
+            )
 
     def min(self, axis=None, skipna=True, *args, **kwargs):
         """

pandas/core/arrays/period.py

@@ -22,7 +22,7 @@
 from pandas.core.dtypes.missing import isna, notna
 
 import pandas.core.algorithms as algos
-from pandas.core.arrays import ExtensionArray, datetimelike as dtl
+from pandas.core.arrays import datetimelike as dtl
 import pandas.core.common as com
 from pandas.core.missing import backfill_1d, pad_1d
 
@@ -92,9 +92,7 @@ def wrapper(self, other):
     return compat.set_function_name(wrapper, opname, cls)
 
 
-class PeriodArray(dtl.DatetimeLikeArrayMixin,
-                  dtl.DatelikeOps,
-                  ExtensionArray):
+class PeriodArray(dtl.DatetimeLikeArrayMixin, dtl.DatelikeOps):
     """
     Pandas ExtensionArray for storing Period data.
 
@@ -418,21 +416,6 @@ def fillna(self, value=None, method=None, limit=None):
             new_values = self.copy()
         return new_values
 
-    def value_counts(self, dropna=False):
-        from pandas import Series, PeriodIndex
-
-        if dropna:
-            values = self[~self.isna()]._data
-        else:
-            values = self._data
-
-        cls = type(self)
-
-        result = algos.value_counts(values, sort=False)
-        index = PeriodIndex(cls(result.index, freq=self.freq),
-                            name=result.index.name)
-        return Series(result.values, index=index, name=result.name)
-
     # --------------------------------------------------------------------
 
     def _time_shift(self, n, freq=None):

pandas/core/arrays/timedeltas.py

@@ -16,7 +16,7 @@
 
 from pandas.core.dtypes.common import (
     _NS_DTYPE, _TD_DTYPE, ensure_int64, is_datetime64_dtype, is_float_dtype,
-    is_integer_dtype, is_list_like, is_object_dtype, is_scalar,
+    is_int64_dtype, is_integer_dtype, is_list_like, is_object_dtype, is_scalar,
     is_string_dtype, is_timedelta64_dtype, is_timedelta64_ns_dtype,
     pandas_dtype)
 from pandas.core.dtypes.dtypes import DatetimeTZDtype
@@ -244,6 +244,16 @@ def _maybe_clear_freq(self):
     # ----------------------------------------------------------------
     # Array-Like / EA-Interface Methods
 
+    def __array__(self, dtype=None):
+        # TODO(https://github.com/pandas-dev/pandas/pull/23593)
+        # Maybe push to parent once datetimetz __array__ is figured out.
+        if is_object_dtype(dtype):
+            return np.array(list(self), dtype=object)
+        elif is_int64_dtype(dtype):
+            return self.asi8
+
+        return self._data
+
     @Appender(dtl.DatetimeLikeArrayMixin._validate_fill_value.__doc__)
     def _validate_fill_value(self, fill_value):
         if isna(fill_value):

pandas/tests/arrays/test_datetimes.py

@@ -123,6 +123,21 @@ def test_repeat_preserves_tz(self):
         expected = DatetimeArray(arr.asi8, freq=None, tz=arr.tz)
         tm.assert_equal(repeated, expected)
 
+    def test_value_counts_preserves_tz(self):
+        dti = pd.date_range('2000', periods=2, freq='D', tz='US/Central')
+        arr = DatetimeArray(dti).repeat([4, 3])
+
+        result = arr.value_counts()
+
+        # Note: not tm.assert_index_equal, since `freq`s do not match
+        assert result.index.equals(dti)
+
+        arr[-2] = pd.NaT
+        result = arr.value_counts()
+        expected = pd.Series([1, 4, 2],
+                             index=[pd.NaT, dti[0], dti[1]])
+        tm.assert_series_equal(result, expected)
+
 
 class TestSequenceToDT64NS(object):