Merge branch 'main' into autofilter-feature

Author: antznette1

doc/source/whatsnew/v3.0.0.rst

@@ -807,6 +807,7 @@ Other Deprecations
 - Deprecated option "future.no_silent_downcasting", as it is no longer used. In a future version accessing this option will raise (:issue:`59502`)
 - Deprecated passing non-Index types to :meth:`Index.join`; explicitly convert to Index first (:issue:`62897`)
 - Deprecated silent casting of non-datetime 'other' to datetime in :meth:`Series.combine_first` (:issue:`62931`)
+- Deprecated silently casting strings to :class:`Timedelta` in binary operations with :class:`Timedelta` (:issue:`59653`)
 - Deprecated slicing on a :class:`Series` or :class:`DataFrame` with a :class:`DatetimeIndex` using a ``datetime.date`` object, explicitly cast to :class:`Timestamp` instead (:issue:`35830`)
 - Deprecated support for the Dataframe Interchange Protocol (:issue:`56732`)
 - Deprecated the 'inplace' keyword from :meth:`Resampler.interpolate`, as passing ``True`` raises ``AttributeError`` (:issue:`58690`)

pandas/_libs/tslibs/timedeltas.pyx

@@ -798,7 +798,7 @@ def _binary_op_method_timedeltalike(op, name):
             return NotImplemented
 
         try:
-            other = Timedelta(other)
+            other = _wrapped_to_timedelta(other)
         except ValueError:
             # failed to parse as timedelta
             return NotImplemented
@@ -2341,7 +2341,7 @@ class Timedelta(_Timedelta):
     def __truediv__(self, other):
         if _should_cast_to_timedelta(other):
             # We interpret NaT as timedelta64("NaT")
-            other = Timedelta(other)
+            other = _wrapped_to_timedelta(other)
             if other is NaT:
                 return np.nan
             if other._creso != self._creso:
@@ -2374,7 +2374,7 @@ class Timedelta(_Timedelta):
     def __rtruediv__(self, other):
         if _should_cast_to_timedelta(other):
             # We interpret NaT as timedelta64("NaT")
-            other = Timedelta(other)
+            other = _wrapped_to_timedelta(other)
             if other is NaT:
                 return np.nan
             if self._creso != other._creso:
@@ -2402,7 +2402,7 @@ class Timedelta(_Timedelta):
         # just defer
         if _should_cast_to_timedelta(other):
             # We interpret NaT as timedelta64("NaT")
-            other = Timedelta(other)
+            other = _wrapped_to_timedelta(other)
             if other is NaT:
                 return np.nan
             if self._creso != other._creso:
@@ -2457,7 +2457,7 @@ class Timedelta(_Timedelta):
         # just defer
         if _should_cast_to_timedelta(other):
             # We interpret NaT as timedelta64("NaT")
-            other = Timedelta(other)
+            other = _wrapped_to_timedelta(other)
             if other is NaT:
                 return np.nan
             if self._creso != other._creso:
@@ -2525,6 +2525,7 @@ def truediv_object_array(ndarray left, ndarray right):
         if cnp.get_timedelta64_value(td64) == NPY_NAT:
             # td here should be interpreted as a td64 NaT
             if _should_cast_to_timedelta(obj):
+                _wrapped_to_timedelta(obj)  # deprecate if allowing string
                 res_value = np.nan
             else:
                 # if its a number then let numpy handle division, otherwise
@@ -2554,6 +2555,7 @@ def floordiv_object_array(ndarray left, ndarray right):
         if cnp.get_timedelta64_value(td64) == NPY_NAT:
             # td here should be interpreted as a td64 NaT
             if _should_cast_to_timedelta(obj):
+                _wrapped_to_timedelta(obj)  # deprecate allowing string
                 res_value = np.nan
             else:
                 # if its a number then let numpy handle division, otherwise
@@ -2585,6 +2587,23 @@ cdef bint is_any_td_scalar(object obj):
     )
 
 
+cdef inline _wrapped_to_timedelta(object other):
+    # Helper for deprecating cases where we cast str to Timedelta
+    td = Timedelta(other)
+    if isinstance(other, str):
+        from pandas.errors import Pandas4Warning
+        warnings.warn(
+            # GH#59653
+            "Scalar operations between Timedelta and string are "
+            "deprecated and will raise in a future version. "
+            "Explicitly cast to Timedelta first.",
+            Pandas4Warning,
+            stacklevel=find_stack_level(),
+        )
+        # When this is enforced, remove str from _should_cast_to_timedelta
+    return td
+
+
 cdef bint _should_cast_to_timedelta(object obj):
     """
     Should we treat this object as a Timedelta for the purpose of a binary op

pandas/conftest.py

@@ -722,7 +722,7 @@ def index(request):
         - ...
     """
     # copy to avoid mutation, e.g. setting .name
-    return indices_dict[request.param].copy()
+    return indices_dict[request.param].copy(deep=False)
 
 
 @pytest.fixture(
@@ -735,7 +735,7 @@ def index_flat(request):
     index fixture, but excluding MultiIndex cases.
     """
     key = request.param
-    return indices_dict[key].copy()
+    return indices_dict[key].copy(deep=False)
 
 
 @pytest.fixture(
@@ -758,18 +758,15 @@ def index_with_missing(request):
 
     MultiIndex is excluded because isna() is not defined for MultiIndex.
     """
-
-    # GH 35538. Use deep copy to avoid illusive bug on np-dev
-    # GHA pipeline that writes into indices_dict despite copy
-    ind = indices_dict[request.param].copy(deep=True)
-    vals = ind.values.copy()
+    ind = indices_dict[request.param]
     if request.param in ["tuples", "mi-with-dt64tz-level", "multi"]:
         # For setting missing values in the top level of MultiIndex
         vals = ind.tolist()
         vals[0] = (None,) + vals[0][1:]
         vals[-1] = (None,) + vals[-1][1:]
         return MultiIndex.from_tuples(vals)
     else:
+        vals = ind.values.copy()
         vals[0] = None
         vals[-1] = None
         return type(ind)(vals)
@@ -850,7 +847,7 @@ def index_or_series_obj(request):
     Fixture for tests on indexes, series and series with a narrow dtype
     copy to avoid mutation, e.g. setting .name
     """
-    return _index_or_series_objs[request.param].copy(deep=True)
+    return _index_or_series_objs[request.param].copy(deep=False)
 
 
 _typ_objects_series = {
@@ -873,7 +870,7 @@ def index_or_series_memory_obj(request):
     series with empty objects type
     copy to avoid mutation, e.g. setting .name
     """
-    return _index_or_series_memory_objs[request.param].copy(deep=True)
+    return _index_or_series_memory_objs[request.param].copy(deep=False)
 
 
 # ----------------------------------------------------------------

pandas/core/frame.py

@@ -341,8 +341,12 @@
       of the left keys.
     * left_anti: use only keys from left frame that are not in right frame, similar
       to SQL left anti join; preserve key order.
+
+      .. versionadded:: 3.0
     * right_anti: use only keys from right frame that are not in left frame, similar
       to SQL right anti join; preserve key order.
+
+      .. versionadded:: 3.0
 on : Hashable or a sequence of the previous
     Column or index level names to join on. These must be found in both
     DataFrames. If `on` is None and not merging on indexes then this defaults
@@ -2927,6 +2931,7 @@ def to_parquet(
         index: bool | None = ...,
         partition_cols: list[str] | None = ...,
         storage_options: StorageOptions = ...,
+        filesystem: Any = ...,
         **kwargs,
     ) -> bytes: ...
 
@@ -2940,6 +2945,7 @@ def to_parquet(
         index: bool | None = ...,
         partition_cols: list[str] | None = ...,
         storage_options: StorageOptions = ...,
+        filesystem: Any = ...,
         **kwargs,
     ) -> None: ...
 
@@ -2953,6 +2959,7 @@ def to_parquet(
         index: bool | None = None,
         partition_cols: list[str] | None = None,
         storage_options: StorageOptions | None = None,
+        filesystem: Any = None,
         **kwargs,
     ) -> bytes | None:
         """
@@ -2992,6 +2999,12 @@ def to_parquet(
             Must be None if path is not a string.
         {storage_options}
 
+        filesystem : fsspec or pyarrow filesystem, default None
+            Filesystem object to use when reading the parquet file. Only implemented
+            for ``engine="pyarrow"``.
+
+            .. versionadded:: 2.1.0
+
         **kwargs
             Additional arguments passed to the parquet library. See
             :ref:`pandas io <io.parquet>` for more details.
@@ -3053,6 +3066,7 @@ def to_parquet(
             index=index,
             partition_cols=partition_cols,
             storage_options=storage_options,
+            filesystem=filesystem,
             **kwargs,
         )