wip

Author: TomAugspurger

pandas/core/accessor.py

@@ -105,7 +105,6 @@ def f(self, *args, **kwargs):
 
 
 def delegate_names(delegate, accessors, typ, overwrite=False):
-    # type (type, list, str, bool) -> Callable
     """
     Add delegated names to a class using a class decorator.
 

pandas/core/arrays/categorical.py

@@ -1259,6 +1259,8 @@ def __array__(self, dtype=None):
             if dtype==None (default), the same dtype as
             categorical.categories.dtype
         """
+        # Need asarray, in case self.categories.values is an ExtensionArray
+        # e.g. in a PeriodIndex. More generally, any Index backed by an EA.
         values = np.asarray(self.categories.values)
         ret = take_1d(values, self._codes)
         if dtype and not is_dtype_equal(dtype, self.categories.dtype):

pandas/core/arrays/datetimelike.py

@@ -11,9 +11,7 @@
 from pandas._libs.tslibs.period import (
     Period, DIFFERENT_FREQ_INDEX, IncompatibleFrequency)
 
-from pandas.errors import (
-    NullFrequencyError, PerformanceWarning, AbstractMethodError
-)
+from pandas.errors import NullFrequencyError, PerformanceWarning
 from pandas import compat
 
 from pandas.tseries import frequencies
@@ -63,15 +61,8 @@ def cmp_method(self, other):
         # comparisons, this will raise in the future
         with warnings.catch_warnings(record=True):
             warnings.filterwarnings("ignore", "elementwise", FutureWarning)
-
-            # XXX: temporary hack till I figure out what's going on.
-            # For PeriodIndex.__eq__, we don't want to convert a scalar
-            # other to a scalar ndarray.
-            if getattr(self, '_wrap_cmp_method', True):
-                other = np.asarray(other)
-
             with np.errstate(all='ignore'):
-                result = op(self.values, other)
+                result = op(self.values, np.asarray(other))
 
         return result
 
@@ -85,10 +76,12 @@ class AttributesMixin(object):
     @property
     def _attributes(self):
         # Inheriting subclass should implement _attributes as a list of strings
+        from pandas.errors import AbstractMethodError
         raise AbstractMethodError(self)
 
     @classmethod
     def _simple_new(cls, values, **kwargs):
+        from pandas.errors import AbstractMethodError
         raise AbstractMethodError(cls)
 
     def _get_attributes_dict(self):
@@ -125,7 +118,7 @@ def _box_func(self):
         """
         box function to get object from internal representation
         """
-        raise AbstractMethodError(self)
+        raise com.AbstractMethodError(self)
 
     def _box_values(self, values):
         """
@@ -358,13 +351,13 @@ def _add_datelike(self, other):
                                 typ=type(other).__name__))
 
     def _sub_datelike(self, other):
-        raise AbstractMethodError(self)
+        raise com.AbstractMethodError(self)
 
     def _sub_period(self, other):
         return NotImplemented
 
     def _add_offset(self, offset):
-        raise AbstractMethodError(self)
+        raise com.AbstractMethodError(self)
 
     def _add_delta(self, other):
         return NotImplemented

pandas/core/arrays/period.py

@@ -22,6 +22,10 @@
     is_float, is_integer, pandas_dtype, is_scalar,
     is_datetime64_dtype,
     is_categorical_dtype,
+    is_object_dtype,
+    is_string_dtype,
+    is_datetime_or_timedelta_dtype,
+    is_dtype_equal,
     ensure_object
 )
 from pandas.core.dtypes.dtypes import PeriodDtype
@@ -68,7 +72,7 @@ def wrapper(self, other):
                 raise IncompatibleFrequency(msg)
 
             result = op(other.ordinal)
-        elif isinstance(other, (ABCPeriodIndex, PeriodArray)):
+        elif isinstance(other, (ABCPeriodIndex, cls)):
             if other.freq != self.freq:
                 msg = DIFFERENT_FREQ_INDEX.format(self.freqstr, other.freqstr)
                 raise IncompatibleFrequency(msg)
@@ -186,7 +190,7 @@ def _complex_new(cls, data=None, ordinal=None, freq=None, start=None,
                                                  freq, fields)
             return cls._from_ordinals(data, freq=freq)
 
-        if isinstance(data, (PeriodArray, PeriodIndex)):
+        if isinstance(data, (cls, PeriodIndex)):
             if freq is None or freq == data.freq:  # no freq change
                 freq = data.freq
                 data = data._ndarray_values
@@ -722,11 +726,36 @@ def astype(self, dtype, copy=True):
         #     super(...), which ends up being... DatetimeIndexOpsMixin?
         # this is complicated.
         # need a pandas_astype(arr, dtype).
-        from pandas.core.arrays import Categorical
-
-        if is_categorical_dtype(dtype):
+        from pandas import Categorical
+
+        dtype = pandas_dtype(dtype)
+
+        if is_object_dtype(dtype):
+            return np.asarray(self, dtype=object)
+        elif is_string_dtype(dtype) and not is_categorical_dtype(dtype):
+            return self._format_native_types()
+        elif is_integer_dtype(dtype):
+            return self.values.astype("i8", copy=copy)
+        elif (is_datetime_or_timedelta_dtype(dtype) and
+              not is_dtype_equal(self.dtype, dtype)) or is_float_dtype(dtype):
+            # disallow conversion between datetime/timedelta,
+            # and conversions for any datetimelike to float
+            msg = 'Cannot cast {name} to dtype {dtype}'
+            raise TypeError(msg.format(name=type(self).__name__, dtype=dtype))
+        elif is_categorical_dtype(dtype):
             return Categorical(self, dtype=dtype)
-        return super(PeriodArray, self).astype(dtype, copy=copy)
+        elif is_period_dtype(dtype):
+            return self.asfreq(dtype.freq)
+        else:
+            return np.asarray(self, dtype=dtype)
+
+    def _box_values_as_index(self):
+        """
+        return object Index which contains boxed values
+        """
+        # This is implemented just for astype
+        from pandas.core.index import Index
+        return Index(self._box_values(self.asi8), dtype=object)
 
 
 PeriodArray._add_comparison_ops()

pandas/core/indexes/base.py

@@ -306,7 +306,9 @@ def __new__(cls, data=None, dtype=None, copy=False, name=None,
               (dtype is not None and is_period_dtype(dtype))):
             from pandas import PeriodIndex
             result = PeriodIndex(data, copy=copy, name=name, **kwargs)
-            if dtype is not None and _o_dtype == dtype:
+            if (dtype is not None and
+                    not is_period_dtype(dtype) and
+                    _o_dtype == dtype):
                 return Index(result.to_pytimedelta(), dtype=_o_dtype)
             else:
                 return result

pandas/core/indexes/period.py

@@ -10,7 +10,6 @@
     is_float,
     is_integer_dtype,
     is_datetime64_any_dtype,
-    is_period_dtype,
     is_bool_dtype,
     pandas_dtype,
 )
@@ -91,11 +90,13 @@ def _delegate_method(self, name, *args, **kwargs):
 )
 @delegate_names(
     PeriodArray,
-    PeriodArray._datetimelike_methods + [
+    [x for x in PeriodArray._datetimelike_methods
+     if x not in {"asfreq", "to_timestamp"}] + [
         '_format_native_types',
         '_maybe_convert_timedelta',
     ],
     "method",
+    # overwrite size, asi8, etc. but not asfreq, to_timestamp
     overwrite=True,
 )
 class PeriodIndex(DatelikeOps, DatetimeIndexOpsMixin,
@@ -445,6 +446,14 @@ def asof_locs(self, where, mask):
 
         return result
 
+    def _box_values_as_index(self):
+        """
+        return object Index which contains boxed values
+        """
+        # TODO(DatetimeArray): remove
+        # Have to add our name.
+        return Index(self._data._box_values_as_index(), name=self.name)
+
     @Appender(_index_shared_docs['astype'])
     def astype(self, dtype, copy=True, how='start'):
         dtype = pandas_dtype(dtype)
@@ -457,16 +466,8 @@ def astype(self, dtype, copy=True, how='start'):
             tz = getattr(dtype, 'tz', None)
             return self.to_timestamp(how=how).tz_localize(tz)
 
-        elif is_integer_dtype(dtype):
-            # astype(int) -> Index, so don't dispatch
-            return self._int64index.copy() if copy else self._int64index
-
-        elif is_period_dtype(dtype):
-            return self.asfreq(freq=dtype.freq)
-
-        return Index(self._data.astype(dtype, copy=copy), name=self.name,
-                     dtype=dtype,  # disable Index inference
-                     copy=False)
+        result = self._data.astype(dtype, copy=copy)
+        return Index(result, name=self.name, dtype=dtype, copy=False)
 
     @Substitution(klass='PeriodIndex')
     @Appender(_shared_docs['searchsorted'])
@@ -499,29 +500,29 @@ def is_full(self):
         values = self.asi8
         return ((values[1:] - values[:-1]) < 2).all()
 
-    # year = _wrap_field_accessor('year')
-    month = _wrap_field_accessor('month')
-    day = _wrap_field_accessor('day')
-    hour = _wrap_field_accessor('hour')
-    minute = _wrap_field_accessor('minute')
-    second = _wrap_field_accessor('second')
-    weekofyear = _wrap_field_accessor('week')
-    week = weekofyear
-    dayofweek = _wrap_field_accessor('dayofweek')
-    weekday = dayofweek
-    dayofyear = day_of_year = _wrap_field_accessor('dayofyear')
-    quarter = _wrap_field_accessor('quarter')
-    qyear = _wrap_field_accessor('qyear')
-    days_in_month = _wrap_field_accessor('days_in_month')
-    daysinmonth = days_in_month
-
-    @property
-    def start_time(self):
-        return self.to_timestamp(how='start')
-
-    @property
-    def end_time(self):
-        return self.to_timestamp(how='end')
+    # # year = _wrap_field_accessor('year')
+    # month = _wrap_field_accessor('month')
+    # day = _wrap_field_accessor('day')
+    # hour = _wrap_field_accessor('hour')
+    # minute = _wrap_field_accessor('minute')
+    # second = _wrap_field_accessor('second')
+    # weekofyear = _wrap_field_accessor('week')
+    # week = weekofyear
+    # dayofweek = _wrap_field_accessor('dayofweek')
+    # weekday = dayofweek
+    # dayofyear = day_of_year = _wrap_field_accessor('dayofyear')
+    # quarter = _wrap_field_accessor('quarter')
+    # qyear = _wrap_field_accessor('qyear')
+    # days_in_month = _wrap_field_accessor('days_in_month')
+    # daysinmonth = days_in_month
+    #
+    # @property
+    # def start_time(self):
+    #     return self.to_timestamp(how='start')
+    #
+    # @property
+    # def end_time(self):
+    #     return self.to_timestamp(how='end')
 
     def _mpl_repr(self):
         # how to represent ourselves to matplotlib

pandas/tests/indexes/period/test_astype.py

@@ -14,7 +14,9 @@ class TestPeriodIndexAsType(object):
     def test_astype_raises(self, dtype):
         # GH#13149, GH#13209
         idx = PeriodIndex(['2016-05-16', 'NaT', NaT, np.NaN], freq='D')
-        msg = 'Cannot cast PeriodIndex to dtype'
+        # XXX: do we care about the name PeriodArray vs. PeriodIndex in the
+        # exception message?
+        msg = 'Cannot cast PeriodArray to dtype'
         with tm.assert_raises_regex(TypeError, msg):
             idx.astype(dtype)