🐛 Series.diff was always setting the dtype to object

Author: Marco Gorelli

doc/source/whatsnew/v1.0.0.rst

@@ -1146,6 +1146,7 @@ ExtensionArray
 - Bug in :class:`arrays.PandasArray` when setting a scalar string (:issue:`28118`, :issue:`28150`).
 - Bug where nullable integers could not be compared to strings (:issue:`28930`)
 - Bug where :class:`DataFrame` constructor raised ``ValueError`` with list-like data and ``dtype`` specified (:issue:`30280`)
+- Bug in :meth:`Series.diff` was always setting the ``dtype`` to ``Object`` (:issue:`30889`)
 
 
 Other

pandas/core/series.py

@@ -2301,6 +2301,19 @@ def diff(self, periods=1):
         dtype: float64
         """
         result = algorithms.diff(com.values_from_object(self), periods)
+        if self.dtype in [
+            "Int8",
+            "Int16",
+            "Int32",
+            "Int64",
+            "UInt8",
+            "UInt16",
+            "UInt32",
+            "UInt64",
+        ]:
+            return self._constructor(
+                result, index=self.index, dtype=self.dtype
+            ).__finalize__(self)
         return self._constructor(result, index=self.index).__finalize__(self)
 
     def autocorr(self, lag=1):

pandas/tests/series/methods/test_diff.py

@@ -3,6 +3,7 @@
 
 from pandas import Series, TimedeltaIndex, date_range
 import pandas._testing as tm
+from pandas.conftest import ALL_EA_INT_DTYPES
 
 
 class TestSeriesDiff:
@@ -75,3 +76,9 @@ def test_diff_object_dtype(self):
         result = s.diff()
         expected = s - s.shift(1)
         tm.assert_series_equal(result, expected)
+
+    @pytest.mark.parametrize("dtype", ALL_EA_INT_DTYPES)
+    def test_nullable_integer(self, dtype):
+        # GH 30889
+        result = Series([1, 2, 3], dtype=dtype).diff().dtype
+        assert result == dtype