From 43eb356224b2ef9112964365bc1a7792278aad69 Mon Sep 17 00:00:00 2001
From: Patrick Hoefler <patrick_hoefler@gmx.net>
Date: Thu, 5 Jan 2023 22:58:17 +0100
Subject: [PATCH 1/5] ENH: Add fast array equal function for indexers

---
 pandas/_libs/lib.pyi          |  8 ++++++++
 pandas/_libs/lib.pyx          | 34 ++++++++++++++++++++++++++++++++++
 pandas/tests/libs/test_lib.py | 21 +++++++++++++++++++++
 3 files changed, 63 insertions(+)

diff --git a/pandas/_libs/lib.pyi b/pandas/_libs/lib.pyi
index 9bc02e90ebb9e..c1e0ef2a22faa 100644
--- a/pandas/_libs/lib.pyi
+++ b/pandas/_libs/lib.pyi
@@ -240,3 +240,11 @@ def get_reverse_indexer(
 ) -> npt.NDArray[np.intp]: ...
 def is_bool_list(obj: list) -> bool: ...
 def dtypes_all_equal(types: list[DtypeObj]) -> bool: ...
+@overload
+def array_equal_fast(
+    left: np.ndarray[np.int64], right: np.ndarray[np.int64]
+) -> bool: ...
+@overload
+def array_equal_fast(
+    left: np.ndarray[np.int32], right: np.ndarray[np.int32]
+) -> bool: ...
diff --git a/pandas/_libs/lib.pyx b/pandas/_libs/lib.pyx
index b56cf2a23a45f..0202e9c1a011b 100644
--- a/pandas/_libs/lib.pyx
+++ b/pandas/_libs/lib.pyx
@@ -50,6 +50,7 @@ from numpy cimport (
     complex128_t,
     flatiter,
     float64_t,
+    int32_t,
     int64_t,
     intp_t,
     ndarray,
@@ -642,6 +643,39 @@ def array_equivalent_object(ndarray left, ndarray right) -> bool:
     return True
 
 
+ctypedef fused int6432_t:
+    int64_t
+    int32_t
+
+
+@cython.wraparound(False)
+@cython.boundscheck(False)
+def array_equal_fast(
+    ndarray[int6432_t, ndim=1] left, ndarray[int6432_t, ndim=1] right,
+) -> bool:
+    """
+    Perform an element by element comparison on 1-d integer arrays, meant for indexer
+    comparisons
+    """
+    cdef:
+        Py_ssize_t i, n = left.size
+
+    if left.size != right.size:
+        return False
+
+    with nogil:
+        for i in range(n):
+
+            if left[i] != right[i]:
+                break
+        else:
+            i = i + 1
+
+    if i != n:
+        return False
+    return True
+
+
 ctypedef fused ndarr_object:
     ndarray[object, ndim=1]
     ndarray[object, ndim=2]
diff --git a/pandas/tests/libs/test_lib.py b/pandas/tests/libs/test_lib.py
index fd7c47d47112f..f3d34c64fa0d6 100644
--- a/pandas/tests/libs/test_lib.py
+++ b/pandas/tests/libs/test_lib.py
@@ -243,6 +243,27 @@ def test_get_reverse_indexer(self):
         expected = np.array([4, 2, 3, 6, 7], dtype=np.intp)
         tm.assert_numpy_array_equal(result, expected)
 
+    @pytest.mark.parametrize("dtype", ["int64", "int32"])
+    def test_array_equal_fast(self, dtype):
+        # GH#
+        left = np.arange(1, 100, dtype=dtype)
+        right = np.arange(1, 100, dtype=dtype)
+        assert lib.array_equal_fast(left, right)
+
+    @pytest.mark.parametrize("dtype", ["int64", "int32"])
+    def test_array_equal_fast_not_equal(self, dtype):
+        # GH#
+        left = np.array([1, 2], dtype=dtype)
+        right = np.array([2, 2], dtype=dtype)
+        assert not lib.array_equal_fast(left, right)
+
+    @pytest.mark.parametrize("dtype", ["int64", "int32"])
+    def test_array_equal_fast_not_equal_shape(self, dtype):
+        # GH#
+        left = np.array([1, 2, 3], dtype=dtype)
+        right = np.array([2, 2], dtype=dtype)
+        assert not lib.array_equal_fast(left, right)
+
 
 def test_cache_readonly_preserve_docstrings():
     # GH18197

From 3ec714e0e4893c4897b535b8728fdd040154f8a9 Mon Sep 17 00:00:00 2001
From: Patrick Hoefler <patrick_hoefler@gmx.net>
Date: Thu, 5 Jan 2023 22:59:04 +0100
Subject: [PATCH 2/5] Add gh ref

---
 pandas/tests/libs/test_lib.py | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/pandas/tests/libs/test_lib.py b/pandas/tests/libs/test_lib.py
index f3d34c64fa0d6..e352250dc748d 100644
--- a/pandas/tests/libs/test_lib.py
+++ b/pandas/tests/libs/test_lib.py
@@ -245,21 +245,21 @@ def test_get_reverse_indexer(self):
 
     @pytest.mark.parametrize("dtype", ["int64", "int32"])
     def test_array_equal_fast(self, dtype):
-        # GH#
+        # GH#50592
         left = np.arange(1, 100, dtype=dtype)
         right = np.arange(1, 100, dtype=dtype)
         assert lib.array_equal_fast(left, right)
 
     @pytest.mark.parametrize("dtype", ["int64", "int32"])
     def test_array_equal_fast_not_equal(self, dtype):
-        # GH#
+        # GH#50592
         left = np.array([1, 2], dtype=dtype)
         right = np.array([2, 2], dtype=dtype)
         assert not lib.array_equal_fast(left, right)
 
     @pytest.mark.parametrize("dtype", ["int64", "int32"])
     def test_array_equal_fast_not_equal_shape(self, dtype):
-        # GH#
+        # GH#50592
         left = np.array([1, 2, 3], dtype=dtype)
         right = np.array([2, 2], dtype=dtype)
         assert not lib.array_equal_fast(left, right)

From c9bf2eaab31e2945836f06bde77c975dff6cecac Mon Sep 17 00:00:00 2001
From: Patrick Hoefler <patrick_hoefler@gmx.net>
Date: Thu, 5 Jan 2023 23:47:10 +0100
Subject: [PATCH 3/5] Fix cython code

---
 pandas/_libs/lib.pyx | 8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)

diff --git a/pandas/_libs/lib.pyx b/pandas/_libs/lib.pyx
index 0202e9c1a011b..208fead7014d6 100644
--- a/pandas/_libs/lib.pyx
+++ b/pandas/_libs/lib.pyx
@@ -659,6 +659,7 @@ def array_equal_fast(
     """
     cdef:
         Py_ssize_t i, n = left.size
+        bint ret = True
 
     if left.size != right.size:
         return False
@@ -667,13 +668,10 @@ def array_equal_fast(
         for i in range(n):
 
             if left[i] != right[i]:
+                ret = False
                 break
-        else:
-            i = i + 1
 
-    if i != n:
-        return False
-    return True
+    return ret
 
 
 ctypedef fused ndarr_object:

From b359e49a6b8591c74fc0cb485cd707b3f2791c1d Mon Sep 17 00:00:00 2001
From: Patrick Hoefler <patrick_hoefler@gmx.net>
Date: Thu, 5 Jan 2023 23:57:51 +0100
Subject: [PATCH 4/5] Remove nogil

---
 pandas/_libs/lib.pyx | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)

diff --git a/pandas/_libs/lib.pyx b/pandas/_libs/lib.pyx
index 208fead7014d6..89e02ac0fa86d 100644
--- a/pandas/_libs/lib.pyx
+++ b/pandas/_libs/lib.pyx
@@ -659,19 +659,16 @@ def array_equal_fast(
     """
     cdef:
         Py_ssize_t i, n = left.size
-        bint ret = True
 
     if left.size != right.size:
         return False
 
-    with nogil:
-        for i in range(n):
+    for i in range(n):
 
-            if left[i] != right[i]:
-                ret = False
-                break
+        if left[i] != right[i]:
+            return False
 
-    return ret
+    return True
 
 
 ctypedef fused ndarr_object:

From 0145e9ad0c2bd63039bee8136264a400af18d527 Mon Sep 17 00:00:00 2001
From: Patrick Hoefler <patrick_hoefler@gmx.net>
Date: Fri, 6 Jan 2023 16:09:46 +0100
Subject: [PATCH 5/5] Fix types

---
 pandas/_libs/lib.pyi | 7 +------
 1 file changed, 1 insertion(+), 6 deletions(-)

diff --git a/pandas/_libs/lib.pyi b/pandas/_libs/lib.pyi
index c1e0ef2a22faa..2439082bf7413 100644
--- a/pandas/_libs/lib.pyi
+++ b/pandas/_libs/lib.pyi
@@ -240,11 +240,6 @@ def get_reverse_indexer(
 ) -> npt.NDArray[np.intp]: ...
 def is_bool_list(obj: list) -> bool: ...
 def dtypes_all_equal(types: list[DtypeObj]) -> bool: ...
-@overload
-def array_equal_fast(
-    left: np.ndarray[np.int64], right: np.ndarray[np.int64]
-) -> bool: ...
-@overload
 def array_equal_fast(
-    left: np.ndarray[np.int32], right: np.ndarray[np.int32]
+    left: np.ndarray, right: np.ndarray  # np.ndarray[np.int64, ndim=1]
 ) -> bool: ...