WIP: change data

Author: TomAugspurger

pandas/core/arrays/datetimelike.py

@@ -43,7 +43,7 @@
 import pandas.core.common as com
 from pandas.core.algorithms import checked_add_with_arr
 
-from .base import ExtensionOpsMixin
+from .base import ExtensionOpsMixin, ExtensionArray
 from pandas.util._decorators import deprecate_kwarg
 
 
@@ -270,7 +270,7 @@ def ceil(self, freq, ambiguous='raise', nonexistent='raise'):
         return self._round(freq, RoundTo.PLUS_INFTY, ambiguous, nonexistent)
 
 
-class DatetimeLikeArrayMixin(DatelikeOps, TimelikeOps, AttributesMixin):
+class DatetimeLikeArrayMixin(DatelikeOps, TimelikeOps, AttributesMixin, ExtensionArray):
     """
     Shared Base/Mixin class for DatetimeArray, TimedeltaArray, PeriodArray
 

pandas/core/arrays/datetimes.py

@@ -209,6 +209,9 @@ def _simple_new(cls, values, freq=None, tz=None, **kwargs):
         return result
 
     def __new__(cls, values, freq=None, tz=None, dtype=None):
+        if isinstance(values, (ABCSeries, ABCIndexClass)):
+            values = values._values
+
         if tz is None and hasattr(values, 'tz'):
             # e.g. DatetimeIndex
             tz = values.tz
@@ -317,7 +320,7 @@ def _generate_range(cls, start, end, periods, freq, tz=None,
         if not right_closed and len(index) and index[-1] == end:
             index = index[:-1]
 
-        return cls._simple_new(index.values, freq=freq, tz=tz)
+        return cls._simple_new(index.asi8, freq=freq, tz=tz)
 
     # -----------------------------------------------------------------
     # Descriptive Properties
@@ -389,12 +392,12 @@ def _resolution(self):
 
     def __array__(self, dtype=None):
         if is_object_dtype(dtype):
+            import pdb; pdb.set_trace
             return np.array(list(self), dtype=object)
         elif is_int64_dtype(dtype):
             return self.asi8
 
-        # TODO: warn that conversion may be lossy?
-        return self._data.view(np.ndarray)  # follow Index.__array__
+        return self._data
 
     def __iter__(self):
         """

pandas/core/indexes/base.py

@@ -687,7 +687,7 @@ def __len__(self):
 
     def __array__(self, dtype=None):
         """ the array interface, return my values """
-        return self._data.view(np.ndarray)
+        return np.asarray(self._data, dtype=dtype)
 
     def __array_wrap__(self, result, context=None):
         """

pandas/core/indexes/datetimelike.py

@@ -25,7 +25,7 @@
 
 from pandas.core import algorithms, ops
 from pandas.core.accessor import PandasDelegate
-from pandas.core.arrays import PeriodArray, ExtensionOpsMixin
+from pandas.core.arrays import ExtensionOpsMixin, PeriodArray
 from pandas.core.arrays.datetimelike import DatetimeLikeArrayMixin
 import pandas.core.indexes.base as ibase
 from pandas.core.indexes.base import Index, _index_shared_docs
@@ -156,6 +156,8 @@ def _box_values_as_index(self):
         """
         return object Index which contains boxed values
         """
+        # XXX: this is broken (not called) for PeriodIndex, which doesn't
+        # define _box_values AFAICT
         from pandas.core.index import Index
         return Index(self._box_values(self.asi8), name=self.name, dtype=object)
 
@@ -705,3 +707,7 @@ def freq(self):
         # Can't simply use delegate_names since our base class is defining
         # freq
         return self._data.freq
+
+    @property
+    def freqstr(self):
+        return self._data.freq.freqstr

pandas/core/indexes/datetimes.py

@@ -13,7 +13,6 @@
 from pandas._libs.tslibs import (
     ccalendar, conversion, fields, parsing, timezones)
 import pandas.compat as compat
-from pandas.core.accessor import delegate_names
 from pandas.util._decorators import Appender, Substitution, cache_readonly
 
 from pandas.core.dtypes.common import (
@@ -25,15 +24,16 @@
 from pandas.core.dtypes.generic import ABCSeries
 from pandas.core.dtypes.missing import isna
 
+from pandas.core.accessor import delegate_names
 from pandas.core.arrays import datetimelike as dtl
 from pandas.core.arrays.datetimes import (
     DatetimeArrayMixin as DatetimeArray, _to_m8)
 from pandas.core.base import _shared_docs
 import pandas.core.common as com
 from pandas.core.indexes.base import Index, _index_shared_docs
 from pandas.core.indexes.datetimelike import (
-    DatetimeIndexOpsMixin, wrap_array_method,
-    wrap_field_accessor, DatelikeIndexMixin, DatetimelikeDelegateMixin)
+    DatelikeIndexMixin, DatetimeIndexOpsMixin, DatetimelikeDelegateMixin,
+    wrap_array_method, wrap_field_accessor)
 from pandas.core.indexes.numeric import Int64Index
 from pandas.core.ops import get_op_result_name
 import pandas.core.tools.datetimes as tools
@@ -185,6 +185,20 @@ class DatetimeIndex(DatelikeIndexMixin,
     """
     _typ = 'datetimeindex'
     _join_precedence = 10
+    # TODO: dispatch
+    @classmethod
+    def _generate_range(cls, start, end, periods, freq, tz=None,
+                        normalize=False, ambiguous="raise",
+                        closed=None):
+        return cls._simple_new(
+            DatetimeArray._generate_range(
+                start, end, periods, freq, tz=tz,
+                normalize=normalize, ambiguous=ambiguous,
+                closed=closed,
+            )
+        )
+    _box_func = DatetimeArray._box_func
+    _box_values = DatetimeArray._box_values
 
     def _join_i8_wrapper(joinf, **kwargs):
         return DatetimeIndexOpsMixin._join_i8_wrapper(joinf, dtype='M8[ns]',
@@ -321,24 +335,23 @@ def _simple_new(cls, values, name=None, freq=None, tz=None,
         if we are passed a non-dtype compat, then coerce using the constructor
         """
         # DatetimeArray._simple_new will accept either i8 or M8[ns] dtypes
-        assert isinstance(values, np.ndarray), type(values)
+        values = DatetimeArray(values, dtype=dtype, freq=freq, tz=tz)
+        # assert isinstance(values, np.ndarray), type(values)
 
         result = super(DatetimeIndex, cls)._simple_new(values, freq, tz,
                                                        **kwargs)
         result.name = name
         result._reset_identity()
         return result
 
+    @property
+    def values(self):
+        return self._data._data
     # --------------------------------------------------------------------
 
     @property
     def _values(self):
-        # tz-naive -> ndarray
-        # tz-aware -> DatetimeIndex
-        if self.tz is not None:
-            return self
-        else:
-            return self.values
+        return self._data
 
     @property
     def tz(self):

pandas/core/indexes/timedeltas.py

@@ -110,6 +110,8 @@ class TimedeltaIndex(TimedeltaDelegateMixin,
 
     _typ = 'timedeltaindex'
     _join_precedence = 10
+    # TODO: dispatch
+    _generate_range = TimedeltaArray._generate_range
 
     def _join_i8_wrapper(joinf, **kwargs):
         return DatetimeIndexOpsMixin._join_i8_wrapper(

pandas/tests/arrays/test_datetimelike.py

@@ -153,7 +153,7 @@ def test_to_period(self, datetime_index, freqstr):
         #  an EA-specific tm.assert_ function
         tm.assert_index_equal(pd.Index(result), pd.Index(expected))
 
-    @pytest.mark.parametrize('propname', pd.DatetimeIndex._bool_ops)
+    @pytest.mark.parametrize('propname', pd.core.arrays.DatetimeArrayMixin._bool_ops)
     def test_bool_properties(self, datetime_index, propname):
         # in this case _bool_ops is just `is_leap_year`
         dti = datetime_index
@@ -165,7 +165,7 @@ def test_bool_properties(self, datetime_index, propname):
 
         tm.assert_numpy_array_equal(result, expected)
 
-    @pytest.mark.parametrize('propname', pd.DatetimeIndex._field_ops)
+    @pytest.mark.parametrize('propname', pd.core.arrays.DatetimeArrayMixin._field_ops)
     def test_int_properties(self, datetime_index, propname):
         dti = datetime_index
         arr = DatetimeArray(dti)
@@ -213,7 +213,7 @@ def test_total_seconds(self, timedelta_index):
 
         tm.assert_numpy_array_equal(result, expected.values)
 
-    @pytest.mark.parametrize('propname', pd.TimedeltaIndex._field_ops)
+    @pytest.mark.parametrize('propname', pd.core.arrays.TimedeltaArrayMixin._field_ops)
     def test_int_properties(self, timedelta_index, propname):
         tdi = timedelta_index
         arr = TimedeltaArray(tdi)