create new test for issue

Author: angela-tarantula

pandas/core/frame.py

@@ -9052,16 +9052,16 @@ def _restore_wide_ints(
                 orig_dt_self = self_original.dtypes.get(col)
                 orig_dt_other = other_original.dtypes.get(col)
 
-                is_at_risk = (orig_dt_self in [np.int64, np.uint64]) or (
+                was_promoted = (orig_dt_self in [np.int64, np.uint64]) or (
                     orig_dt_other in [np.int64, np.uint64]
                 )
 
-                if is_at_risk and not isna(ser).any():
+                if was_promoted and not isna(ser).any():
                     dtypes_to_resolve = [
                         dt for dt in (orig_dt_self, orig_dt_other) if dt is not None
                     ]
                     if dtypes_to_resolve:
-                        # if we had different dtypes, possibly promote
+                        # if we had different dtypes, reconcile
                         cast_map[col] = find_common_type(dtypes_to_resolve)
 
             if cast_map:

pandas/tests/frame/methods/test_combine_first.py

@@ -398,6 +398,21 @@ def test_combine_first_string_dtype_only_na(self, nullable_string_dtype):
         ).set_index(["a", "b"])
         tm.assert_frame_equal(result, expected)
 
+    @pytest.mark.parametrize(
+        "wide_val, dtype, EAdtype",
+        (
+            (1666880195890293744, "uint64", "UInt64"),
+            (-1666880195890293744, "int64", "Int64"),
+        ),
+    )
+    def test_combine_first_preserve_precision(self, wide_val, dtype, EAdtype):
+        # GH#60128
+        df1 = DataFrame({"A": [wide_val, 5]}, dtype=dtype)
+        df2 = DataFrame({"A": [6, 7, wide_val]}, dtype=dtype)
+        result = df1.combine_first(df2)
+        expected = DataFrame({"A": [wide_val, 5, wide_val]}, dtype=dtype)
+        tm.assert_frame_equal(result, expected)
+
 
 @pytest.mark.parametrize(
     "scalar1, scalar2",