DEPR: invalid fill_value in .shift

doc/source/whatsnew/v3.0.0.rst

@@ -660,6 +660,7 @@ Other Deprecations
 - Deprecated the ``arg`` parameter of ``Series.map``; pass the added ``func`` argument instead. (:issue:`61260`)
 - Deprecated using ``epoch`` date format in :meth:`DataFrame.to_json` and :meth:`Series.to_json`, use ``iso`` instead. (:issue:`57063`)
 - Deprecated allowing ``fill_value`` that cannot be held in the original dtype (excepting NA values for integer and bool dtypes) in :meth:`Series.unstack` and :meth:`DataFrame.unstack` (:issue:`12189`, :issue:`53868`)
+- Deprecated allowing ``fill_value`` that cannot be held in the original dtype (excepting NA values for integer and bool dtypes) in :meth:`Series.shift` and :meth:`DataFrame.shift` (:issue:`53802`)
 
 .. ---------------------------------------------------------------------------
 .. _whatsnew_300.prior_deprecations:

pandas/core/internals/blocks.py

@@ -1449,6 +1449,18 @@ def shift(self, periods: int, fill_value: Any = None) -> list[Block]:
                 fill_value,
             )
         except LossySetitemError:
+            if self.dtype.kind not in "iub" or not is_valid_na_for_dtype(
+                fill_value, self.dtype
+            ):
+                # GH#53802
+                warnings.warn(
+                    "shifting with a fill value that cannot be held by "
+                    "original dtype is deprecated and will raise in a future "
+                    "version. Explicitly cast to the desired dtype before "
+                    "shifting instead.",
+                    Pandas4Warning,
+                    stacklevel=find_stack_level(),
+                )
             nb = self.coerce_to_target_dtype(fill_value, raise_on_upcast=False)
             return nb.shift(periods, fill_value=fill_value)
 

pandas/tests/frame/methods/test_shift.py

@@ -1,6 +1,8 @@
 import numpy as np
 import pytest
 
+from pandas.errors import Pandas4Warning
+
 import pandas as pd
 from pandas import (
     CategoricalIndex,
@@ -766,3 +768,29 @@ def test_series_shift_interval_preserves_closed(self):
         result = ser.shift(1)
         expected = Series([np.nan, pd.Interval(1, 2, closed="right")])
         tm.assert_series_equal(result, expected)
+
+    def test_shift_invalid_fill_value_deprecation(self):
+        # GH#53802
+        df = DataFrame(
+            {
+                "a": [1, 2, 3],
+                "b": [True, False, True],
+            }
+        )
+
+        msg = "shifting with a fill value that cannot"
+        with tm.assert_produces_warning(Pandas4Warning, match=msg):
+            df.shift(1, fill_value="foo")
+
+        with tm.assert_produces_warning(Pandas4Warning, match=msg):
+            df["a"].shift(1, fill_value="foo")
+        with tm.assert_produces_warning(Pandas4Warning, match=msg):
+            df["b"].shift(1, fill_value="foo")
+
+        # An incompatible null value
+        with tm.assert_produces_warning(Pandas4Warning, match=msg):
+            df.shift(1, fill_value=NaT)
+        with tm.assert_produces_warning(Pandas4Warning, match=msg):
+            df["a"].shift(1, fill_value=NaT)
+        with tm.assert_produces_warning(Pandas4Warning, match=msg):
+            df["b"].shift(1, fill_value=NaT)