From 4ce6674641def5a68dce633d3a21f17438ae48de Mon Sep 17 00:00:00 2001
From: GALI PREM SAGAR <sagarprem75@gmail.com>
Date: Mon, 6 May 2024 14:43:20 -0500
Subject: [PATCH] Return `int64` when pandas compatible mode is turned on for
 `get_indexer` (#15659)

Fixes: #15658

This PR makes a change to `get_indexer` to return `int64` indices when pandas compatible mode is turned on.

Forks out of https://github.com/rapidsai/cudf/pull/14534

Authors:
  - GALI PREM SAGAR (https://github.com/galipremsagar)

Approvers:
  - Matthew Roeschke (https://github.com/mroeschke)

URL: https://github.com/rapidsai/cudf/pull/15659
---
 python/cudf/cudf/core/_base_index.py |  6 ++++++
 python/cudf/cudf/core/index.py       |  8 ++++----
 python/cudf/cudf/core/multiindex.py  |  7 ++++---
 python/cudf/cudf/tests/test_index.py | 20 ++++++++++++++++++++
 4 files changed, 34 insertions(+), 7 deletions(-)

diff --git a/python/cudf/cudf/core/_base_index.py b/python/cudf/cudf/core/_base_index.py
index b5630ff9a54..fe0f39f9d0a 100644
--- a/python/cudf/cudf/core/_base_index.py
+++ b/python/cudf/cudf/core/_base_index.py
@@ -2205,3 +2205,9 @@ def _split(self, splits):
 
 def _get_result_name(left_name, right_name):
     return left_name if _is_same_name(left_name, right_name) else None
+
+
+def _return_get_indexer_result(result):
+    if cudf.get_option("mode.pandas_compatible"):
+        return result.astype("int64")
+    return result
diff --git a/python/cudf/cudf/core/index.py b/python/cudf/cudf/core/index.py
index b51751a1b55..a2ad10a0590 100644
--- a/python/cudf/cudf/core/index.py
+++ b/python/cudf/cudf/core/index.py
@@ -38,7 +38,7 @@
     is_list_like,
     is_scalar,
 )
-from cudf.core._base_index import BaseIndex
+from cudf.core._base_index import BaseIndex, _return_get_indexer_result
 from cudf.core._compat import PANDAS_LT_300
 from cudf.core.column import (
     CategoricalColumn,
@@ -1256,11 +1256,11 @@ def get_indexer(self, target, method=None, limit=None, tolerance=None):
         )
 
         if not len(self):
-            return result.values
+            return _return_get_indexer_result(result.values)
         try:
             lcol, rcol = _match_join_keys(needle, self._column, "inner")
         except ValueError:
-            return result.values
+            return _return_get_indexer_result(result.values)
 
         scatter_map, indices = libcudf.join.join([lcol], [rcol], how="inner")
         (result,) = libcudf.copying.scatter([indices], scatter_map, [result])
@@ -1287,7 +1287,7 @@ def get_indexer(self, target, method=None, limit=None, tolerance=None):
                 "{['ffill'/'pad', 'bfill'/'backfill', 'nearest', None]}"
             )
 
-        return result_series.to_cupy()
+        return _return_get_indexer_result(result_series.to_cupy())
 
     @_cudf_nvtx_annotate
     def get_loc(self, key):
diff --git a/python/cudf/cudf/core/multiindex.py b/python/cudf/cudf/core/multiindex.py
index 1ab42df111f..c3184f51a4c 100644
--- a/python/cudf/cudf/core/multiindex.py
+++ b/python/cudf/cudf/core/multiindex.py
@@ -23,6 +23,7 @@
 from cudf.api.extensions import no_default
 from cudf.api.types import is_integer, is_list_like, is_object_dtype
 from cudf.core import column
+from cudf.core._base_index import _return_get_indexer_result
 from cudf.core.frame import Frame
 from cudf.core.index import (
     BaseIndex,
@@ -1858,11 +1859,11 @@ def get_indexer(self, target, method=None, limit=None, tolerance=None):
             dtype=libcudf.types.size_type_dtype,
         )
         if not len(self):
-            return result.values
+            return _return_get_indexer_result(result.values)
         try:
             target = cudf.MultiIndex.from_tuples(target)
         except TypeError:
-            return result.values
+            return _return_get_indexer_result(result.values)
 
         join_keys = [
             _match_join_keys(lcol, rcol, "inner")
@@ -1892,7 +1893,7 @@ def get_indexer(self, target, method=None, limit=None, tolerance=None):
                 "{['ffill'/'pad', 'bfill'/'backfill', None]}"
             )
 
-        return result_series.to_cupy()
+        return _return_get_indexer_result(result_series.to_cupy())
 
     @_cudf_nvtx_annotate
     def get_loc(self, key):
diff --git a/python/cudf/cudf/tests/test_index.py b/python/cudf/cudf/tests/test_index.py
index 104a5fc0ffa..4ff1beb0a9a 100644
--- a/python/cudf/cudf/tests/test_index.py
+++ b/python/cudf/cudf/tests/test_index.py
@@ -1741,6 +1741,10 @@ def test_get_indexer_single_unique_numeric(idx, key, method):
 
         assert_eq(expected, got)
 
+        with cudf.option_context("mode.pandas_compatible", True):
+            got = gi.get_indexer(key, method=method)
+        assert_eq(expected, got, check_dtype=True)
+
 
 @pytest.mark.parametrize(
     "idx",
@@ -1770,6 +1774,12 @@ def test_get_indexer_rangeindex(idx, key, method, tolerance):
 
     assert_eq(expected, got)
 
+    with cudf.option_context("mode.pandas_compatible", True):
+        got = gi.get_indexer(
+            key, method=method, tolerance=None if method is None else tolerance
+        )
+    assert_eq(expected, got, check_dtype=True)
+
 
 @pytest.mark.parametrize(
     "idx",
@@ -1950,6 +1960,11 @@ def test_get_indexer_single_duplicate_string(idx, key, method):
 
         assert_eq(expected, got)
 
+        with cudf.option_context("mode.pandas_compatible", True):
+            got = gi.get_indexer(key, method=method)
+
+        assert_eq(expected, got, check_dtype=True)
+
 
 @pytest.mark.parametrize(
     "idx",
@@ -2009,6 +2024,11 @@ def test_get_indexer_multi_numeric(idx, key, method):
 
     assert_eq(expected, got)
 
+    with cudf.option_context("mode.pandas_compatible", True):
+        got = gi.get_indexer(key, method=method)
+
+    assert_eq(expected, got, check_dtype=True)
+
 
 @pytest.mark.parametrize(
     "idx",