performance improved approach

Author: divya1974

pandas/core/algorithms.py

@@ -579,6 +579,35 @@ def f(c, v):
             f = lambda a, b: np.isin(a, b).ravel()
 
     else:
+        # Fast-path for integer mixes: if both sides are integer-kind and
+        # have different dtypes, avoid upcasting to float64 (which loses
+        # precision for large 64-bit integers). When possible, perform the
+        # comparison in unsigned 64-bit space which preserves exact integer
+        # equality and uses the integer hashtable for performance.
+        if (
+            values.dtype.kind in "iu"
+            and comps_array.dtype.kind in "iu"
+            and not is_dtype_equal(values.dtype, comps_array.dtype)
+        ):
+            try:
+                # only proceed when both arrays are non-empty
+                if values.size > 0 and comps_array.size > 0:
+                    signed_negative = False
+                    if values.dtype.kind == "i":
+                        # using min is vectorized and fast
+                        signed_negative = values.min() < 0
+                    if comps_array.dtype.kind == "i":
+                        signed_negative = signed_negative or (comps_array.min() < 0)
+
+                    if not signed_negative:
+                        # safe to cast both to uint64 for exact comparison
+                        values_u = values.astype("uint64", copy=False)
+                        comps_u = comps_array.astype("uint64", copy=False)
+                        return htable.ismember(comps_u, values_u)
+            except Exception:
+                # fall back to generic behavior on any error
+                pass
+
         common = np_find_common_type(values.dtype, comps_array.dtype)
         values = values.astype(common, copy=False)
         comps_array = comps_array.astype(common, copy=False)

pandas/tests/series/methods/test_isin.py

@@ -267,3 +267,13 @@ def test_isin_filtering_on_iterable(data, isin):
     expected_result = Series([True, True, False])
 
     tm.assert_series_equal(result, expected_result)
+
+
+def test_isin_int64_vs_uint64_mismatch():
+    # Regression test for mixing signed int64 Series with uint64 values
+    ser = Series([1378774140726870442], dtype=np.int64)
+    vals = [np.uint64(1378774140726870528)]
+
+    res = ser.isin(vals)
+    expected = Series([False])
+    tm.assert_series_equal(res, expected)