Merge branch 'main' into daydst2

Author: jbrockmendel

doc/source/whatsnew/v2.2.0.rst

@@ -234,6 +234,7 @@ For example:
 Other Deprecations
 ^^^^^^^^^^^^^^^^^^
 - Changed :meth:`Timedelta.resolution_string` to return ``h``, ``min``, ``s``, ``ms``, ``us``, and ``ns`` instead of ``H``, ``T``, ``S``, ``L``, ``U``, and ``N``, for compatibility with respective deprecations in frequency aliases (:issue:`52536`)
+- Deprecated :meth:`Index.format`, use ``index.astype(str)`` or ``index.map(formatter)`` instead (:issue:`55413`)
 - Deprecated allowing non-keyword arguments in :meth:`DataFrame.to_clipboard`. (:issue:`54229`)
 - Deprecated allowing non-keyword arguments in :meth:`DataFrame.to_csv` except ``path_or_buf``. (:issue:`54229`)
 - Deprecated allowing non-keyword arguments in :meth:`DataFrame.to_dict`. (:issue:`54229`)
@@ -261,6 +262,7 @@ Other Deprecations
 - Deprecated the extension test classes ``BaseNoReduceTests``, ``BaseBooleanReduceTests``, and ``BaseNumericReduceTests``, use ``BaseReduceTests`` instead (:issue:`54663`)
 - Deprecated the option ``mode.data_manager`` and the ``ArrayManager``; only the ``BlockManager`` will be available in future versions (:issue:`55043`)
 - Deprecating downcasting the results of :meth:`DataFrame.fillna`, :meth:`Series.fillna`, :meth:`DataFrame.ffill`, :meth:`Series.ffill`, :meth:`DataFrame.bfill`, :meth:`Series.bfill` in object-dtype cases. To opt in to the future version, use ``pd.set_option("future.no_silent_downcasting", True)`` (:issue:`54261`)
+-
 
 .. ---------------------------------------------------------------------------
 .. _whatsnew_220.performance:
@@ -285,17 +287,20 @@ Bug fixes
 Categorical
 ^^^^^^^^^^^
 - :meth:`Categorical.isin` raising ``InvalidIndexError`` for categorical containing overlapping :class:`Interval` values (:issue:`34974`)
+- Bug in :meth:`CategoricalDtype.__eq__` returning false for unordered categorical data with mixed types (:issue:`55468`)
 -
 
 Datetimelike
 ^^^^^^^^^^^^
 - Bug in :meth:`DatetimeIndex.union` returning object dtype for tz-aware indexes with the same timezone but different units (:issue:`55238`)
--
+- Bug in :meth:`Tick.delta` with very large ticks raising ``OverflowError`` instead of ``OutOfBoundsTimedelta`` (:issue:`55503`)
+- Bug in addition or subtraction of very large :class:`Tick` objects with :class:`Timestamp` or :class:`Timedelta` objects raising ``OverflowError`` instead of ``OutOfBoundsTimedelta`` (:issue:`55503`)
+
 
 Timedelta
 ^^^^^^^^^
+- Bug in :class:`Timedelta` construction raising ``OverflowError`` instead of ``OutOfBoundsTimedelta`` (:issue:`55503`)
 - Bug in rendering (``__repr__``) of :class:`TimedeltaIndex` and :class:`Series` with timedelta64 values with non-nanosecond resolution entries that are all multiples of 24 hours failing to use the compact representation used in the nanosecond cases (:issue:`55405`)
--
 
 Timezones
 ^^^^^^^^^
@@ -352,7 +357,7 @@ I/O
 
 Period
 ^^^^^^
--
+- Bug in :class:`Period` addition silently wrapping around instead of raising ``OverflowError`` (:issue:`55503`)
 -
 
 Plotting

pandas/_libs/tslibs/offsets.pyx

@@ -964,7 +964,12 @@ cdef class Tick(SingleConstructorOffset):
 
     @property
     def delta(self):
-        return self.n * Timedelta(self._nanos_inc)
+        try:
+            return self.n * Timedelta(self._nanos_inc)
+        except OverflowError as err:
+            # GH#55503 as_unit will raise a more useful OutOfBoundsTimedelta
+            Timedelta(self).as_unit("ns")
+            raise AssertionError("This should not be reached.")
 
     @property
     def nanos(self) -> int64_t:

pandas/_libs/tslibs/period.pyx

@@ -1815,7 +1815,7 @@ cdef class _Period(PeriodMixin):
 
     def _add_timedeltalike_scalar(self, other) -> "Period":
         cdef:
-            int64_t inc
+            int64_t inc, ordinal
 
         if not self._dtype._is_tick_like():
             raise IncompatibleFrequency("Input cannot be converted to "
@@ -1837,8 +1837,8 @@ cdef class _Period(PeriodMixin):
         except ValueError as err:
             raise IncompatibleFrequency("Input cannot be converted to "
                                         f"Period(freq={self.freqstr})") from err
-        # TODO: overflow-check here
-        ordinal = self.ordinal + inc
+        with cython.overflowcheck(True):
+            ordinal = self.ordinal + inc
         return Period(ordinal=ordinal, freq=self.freq)
 
     def _add_offset(self, other) -> "Period":
@@ -1851,6 +1851,7 @@ cdef class _Period(PeriodMixin):
         ordinal = self.ordinal + other.n
         return Period(ordinal=ordinal, freq=self.freq)
 
+    @cython.overflowcheck(True)
     def __add__(self, other):
         if not is_period_object(self):
             # cython semantics; this is analogous to a call to __radd__

pandas/_libs/tslibs/timedeltas.pyx

@@ -1784,7 +1784,7 @@ class Timedelta(_Timedelta):
                 )
 
             # GH43764, convert any input to nanoseconds first and then
-            # create the timestamp. This ensures that any potential
+            # create the timedelta. This ensures that any potential
             # nanosecond contributions from kwargs parsed as floats
             # are taken into consideration.
             seconds = int((
@@ -1797,12 +1797,24 @@ class Timedelta(_Timedelta):
                 ) * 1_000_000_000
             )
 
-            value = np.timedelta64(
-                int(kwargs.get("nanoseconds", 0))
-                + int(kwargs.get("microseconds", 0) * 1_000)
-                + int(kwargs.get("milliseconds", 0) * 1_000_000)
-                + seconds
-            )
+            ns = kwargs.get("nanoseconds", 0)
+            us = kwargs.get("microseconds", 0)
+            ms = kwargs.get("milliseconds", 0)
+            try:
+                value = np.timedelta64(
+                    int(ns)
+                    + int(us * 1_000)
+                    + int(ms * 1_000_000)
+                    + seconds
+                )
+            except OverflowError as err:
+                # GH#55503
+                msg = (
+                    f"seconds={seconds}, milliseconds={ms}, "
+                    f"microseconds={us}, nanoseconds={ns}"
+                )
+                raise OutOfBoundsTimedelta(msg) from err
+
         if unit in {"Y", "y", "M"}:
             raise ValueError(
                 "Units 'M', 'Y', and 'y' are no longer supported, as they do not "

pandas/_libs/tslibs/tzconversion.pyx

@@ -425,7 +425,11 @@ timedelta-like}
     return result.base  # .base to get underlying ndarray
 
 
-cdef Py_ssize_t bisect_right_i8(int64_t *data, int64_t val, Py_ssize_t n):
+cdef Py_ssize_t bisect_right_i8(
+    const int64_t *data,
+    int64_t val,
+    Py_ssize_t n
+) noexcept:
     # Caller is responsible for checking n > 0
     # This looks very similar to local_search_right in the ndarray.searchsorted
     #  implementation.
@@ -463,7 +467,7 @@ cdef str _render_tstamp(int64_t val, NPY_DATETIMEUNIT creso):
 
 cdef _get_utc_bounds(
     ndarray[int64_t] vals,
-    int64_t* tdata,
+    const int64_t* tdata,
     Py_ssize_t ntrans,
     const int64_t[::1] deltas,
     NPY_DATETIMEUNIT creso,

pandas/core/dtypes/dtypes.py

@@ -456,7 +456,7 @@ def __eq__(self, other: object) -> bool:
 
             # With object-dtype we need a comparison that identifies
             #  e.g. int(2) as distinct from float(2)
-            return hash(self) == hash(other)
+            return set(left) == set(right)
 
     def __repr__(self) -> str_type:
         if self.categories is None: