Make _freq/freq/tz/_tz/dtype/_dtype/offset/_offset all inherit reliably (#24517)

Author: jbrockmendel

pandas/core/arrays/datetimelike.py

@@ -1138,8 +1138,6 @@ def _time_shift(self, periods, freq=None):
  freq = frequencies.to_offset(freq)
  offset = periods * freq
  result = self + offset
- if hasattr(self, 'tz'):
- result._tz = self.tz
  return result
 
  if periods == 0:

pandas/core/arrays/datetimes.py

@@ -210,7 +210,7 @@ class DatetimeArrayMixin(dtl.DatetimeLikeArrayMixin,
  # Constructors
 
  _attributes = ["freq", "tz"]
- _tz = None
+ _dtype = None # type: Union[np.dtype, DatetimeTZDtype]
  _freq = None
 
  @classmethod
@@ -231,8 +231,13 @@ def _simple_new(cls, values, freq=None, tz=None):
  result = object.__new__(cls)
  result._data = values
  result._freq = freq
- tz = timezones.maybe_get_tz(tz)
- result._tz = timezones.tz_standardize(tz)
+ if tz is None:
+ dtype = _NS_DTYPE
+ else:
+ tz = timezones.maybe_get_tz(tz)
+ tz = timezones.tz_standardize(tz)
+ dtype = DatetimeTZDtype('ns', tz)
+ result._dtype = dtype
  return result
 
  def __new__(cls, values, freq=None, tz=None, dtype=None, copy=False,
@@ -399,9 +404,7 @@ def dtype(self):
  If the values are tz-aware, then the ``DatetimeTZDtype``
  is returned.
  """
- if self.tz is None:
- return _NS_DTYPE
- return DatetimeTZDtype('ns', self.tz)
+ return self._dtype
 
  @property
  def tz(self):
@@ -411,10 +414,10 @@ def tz(self):
  Returns
  -------
  datetime.tzinfo, pytz.tzinfo.BaseTZInfo, dateutil.tz.tz.tzfile, or None
-  Returns None when the array is tz-naive.
+ Returns None when the array is tz-naive.
  """
  # GH 18595
- return self._tz
+ return getattr(self._dtype, "tz", None)
 
  @tz.setter
  def tz(self, value):

pandas/core/indexes/datetimelike.py

@@ -70,6 +70,15 @@ class DatetimeIndexOpsMixin(ExtensionOpsMixin):
  _maybe_mask_results = ea_passthrough("_maybe_mask_results")
  __iter__ = ea_passthrough("__iter__")
 
+ @property
+ def freq(self):
+ return self._eadata.freq
+
+ @freq.setter
+ def freq(self, value):
+ # validation is handled by _eadata setter
+ self._eadata.freq = value
+
  @property
  def freqstr(self):
  return self._eadata.freqstr
@@ -98,6 +107,10 @@ def wrapper(self, other):
  wrapper.__name__ = '__{}__'.format(op.__name__)
  return wrapper
 
+ @property
+ def _ndarray_values(self):
+ return self._eadata._ndarray_values
+
  # ------------------------------------------------------------------------
 
  def equals(self, other):

pandas/core/indexes/datetimes.py

@@ -321,11 +321,10 @@ def _simple_new(cls, values, name=None, freq=None, tz=None, dtype=None):
 
  dtarr = DatetimeArray._simple_new(values, freq=freq, tz=tz)
  result = object.__new__(cls)
- result._data = dtarr._data
- result._freq = dtarr.freq
- result._tz = dtarr.tz
+ result._eadata = dtarr
  result.name = name
  # For groupby perf. See note in indexes/base about _index_data
+ # TODO: make sure this is updated correctly if edited
  result._index_data = result._data
  result._reset_identity()
  return result
@@ -345,19 +344,6 @@ def _values(self):
  else:
  return self.values
 
- @property
- def tz(self):
- # GH 18595
- return self._tz
-
- @tz.setter
- def tz(self, value):
- # GH 3746: Prevent localizing or converting the index by setting tz
- raise AttributeError("Cannot directly set timezone. Use tz_localize() "
- "or tz_convert() as appropriate")
-
- tzinfo = tz
-
  @property
  def size(self):
  # TODO: Remove this when we have a DatetimeTZArray
@@ -416,15 +402,18 @@ def __setstate__(self, state):
  data = np.empty(nd_state[1], dtype=nd_state[2])
  np.ndarray.__setstate__(data, nd_state)
 
+ freq = own_state[1]
+ tz = timezones.tz_standardize(own_state[2])
+ dtarr = DatetimeArray._simple_new(data, freq=freq, tz=tz)
+
  self.name = own_state[0]
- self._freq = own_state[1]
- self._tz = timezones.tz_standardize(own_state[2])
 
  else: # pragma: no cover
  data = np.empty(state)
  np.ndarray.__setstate__(data, state)
+ dtarr = DatetimeArray(data)
 
- self._data = data
+ self._eadata = dtarr
  self._reset_identity()
 
  else:
@@ -502,7 +491,9 @@ def union(self, other):
  else:
  result = Index.union(this, other)
  if isinstance(result, DatetimeIndex):
- result._tz = timezones.tz_standardize(this.tz)
+ # TODO: we shouldn't be setting attributes like this;
+ # in all the tests this equality already holds
+ result._eadata._dtype = this.dtype
  if (result.freq is None and
  (this.freq is not None or other.freq is not None)):
  result.freq = to_offset(result.inferred_freq)
@@ -530,11 +521,12 @@ def union_many(self, others):
  if this._can_fast_union(other):
  this = this._fast_union(other)
  else:
- tz = this.tz
+ dtype = this.dtype
  this = Index.union(this, other)
  if isinstance(this, DatetimeIndex):
- this._tz = timezones.tz_standardize(tz)
-
+ # TODO: we shouldn't be setting attributes like this;
+ # in all the tests this equality already holds
+ this._eadata._dtype = dtype
  return this
 
  def _can_fast_union(self, other):
@@ -1129,9 +1121,20 @@ def slice_indexer(self, start=None, end=None, step=None, kind=None):
  # Wrapping DatetimeArray
 
  @property
- def _eadata(self):
- return DatetimeArray._simple_new(self._data,
- tz=self.tz, freq=self.freq)
+ def _data(self):
+ return self._eadata._data
+
+ @property
+ def tz(self):
+ # GH#18595
+ return self._eadata.tz
+
+ @tz.setter
+ def tz(self, value):
+ # GH#3746; DatetimeArray will raise to disallow setting
+ self._eadata.tz = value
+
+ tzinfo = tz
 
  # Compat for frequency inference, see GH#23789
  _is_monotonic_increasing = Index.is_monotonic_increasing
@@ -1168,18 +1171,6 @@ def offset(self, value):
  warnings.warn(msg, FutureWarning, stacklevel=2)
  self.freq = value
 
- @property
- def freq(self):
- return self._freq
-
- @freq.setter
- def freq(self, value):
- if value is not None:
- # let DatetimeArray to validation
- self._eadata.freq = value
-
- self._freq = to_offset(value)
-
  def __getitem__(self, key):
  result = self._eadata.__getitem__(key)
  if is_scalar(result):

pandas/core/indexes/period.py

@@ -8,8 +8,7 @@
 from pandas._libs.tslibs import NaT, iNaT, resolution
 from pandas._libs.tslibs.period import (
  DIFFERENT_FREQ, IncompatibleFrequency, Period)
-from pandas.util._decorators import (
- Appender, Substitution, cache_readonly, deprecate_kwarg)
+from pandas.util._decorators import Appender, Substitution, cache_readonly
 
 from pandas.core.dtypes.common import (
  is_bool_dtype, is_datetime64_any_dtype, is_float, is_float_dtype,
@@ -288,10 +287,6 @@ def _simple_new(cls, values, name=None, freq=None, **kwargs):
  def _eadata(self):
  return self._data
 
- @property
- def _ndarray_values(self):
- return self._data._ndarray_values
-
  @property
  def values(self):
  return np.asarray(self)
@@ -472,34 +467,6 @@ def _int64index(self):
  # ------------------------------------------------------------------------
  # Index Methods
 
- @deprecate_kwarg(old_arg_name='n', new_arg_name='periods')
- def shift(self, periods):
- """
- Shift index by desired number of increments.
-
- This method is for shifting the values of period indexes
- by a specified time increment.
-
- Parameters
- ----------
- periods : int, default 1
- Number of periods (or increments) to shift by,
- can be positive or negative.
-
- .. versionchanged:: 0.24.0
-
- Returns
- -------
- pandas.PeriodIndex
- Shifted index.
-
- See Also
- --------
- DatetimeIndex.shift : Shift values of DatetimeIndex.
- """
- i8values = self._data._time_shift(periods)
- return self._simple_new(i8values, name=self.name, freq=self.freq)
-
  def _coerce_scalar_to_index(self, item):
  """
  we need to coerce a scalar to a compat for our index type

pandas/core/indexes/timedeltas.py

@@ -235,11 +235,12 @@ def _simple_new(cls, values, name=None, freq=None, dtype=_TD_DTYPE):
  freq = to_offset(freq)
  tdarr = TimedeltaArray._simple_new(values, freq=freq)
  result = object.__new__(cls)
- result._data = tdarr._data
- result._freq = tdarr._freq
+ result._eadata = tdarr
  result.name = name
  # For groupby perf. See note in indexes/base about _index_data
- result._index_data = result._data
+ # TODO: make sure this is updated correctly if edited
+ result._index_data = tdarr._data
+
  result._reset_identity()
  return result
 
@@ -279,8 +280,8 @@ def _format_native_types(self, na_rep='NaT', date_format=None, **kwargs):
  # Wrapping TimedeltaArray
 
  @property
- def _eadata(self):
- return TimedeltaArray._simple_new(self._data, freq=self.freq)
+ def _data(self):
+ return self._eadata._data
 
  __mul__ = _make_wrapped_arith_op("__mul__")
  __rmul__ = _make_wrapped_arith_op("__rmul__")
@@ -316,18 +317,6 @@ def __getitem__(self, key):
  return result
  return type(self)(result, name=self.name)
 
- @property
- def freq(self): # TODO: get via eadata
- return self._freq
-
- @freq.setter
- def freq(self, value): # TODO: get via eadata
- if value is not None:
- # dispatch to TimedeltaArray to validate frequency
- self._eadata.freq = value
-
- self._freq = to_offset(value)
-
  # -------------------------------------------------------------------
 
  @Appender(_index_shared_docs['astype'])

pandas/tests/arithmetic/test_timedelta64.py

@@ -1476,7 +1476,7 @@ def test_tdi_rmul_arraylike(self, other, box_with_array):
 
  tdi = TimedeltaIndex(['1 Day'] * 10)
  expected = timedelta_range('1 days', '10 days')
- expected._freq = None
+ expected._eadata._freq = None
 
  tdi = tm.box_expected(tdi, box)
  expected = tm.box_expected(expected, xbox)