Merge branch-25.06 into branch-25.08

Author: Matt711

.github/workflows/pr.yaml

@@ -43,7 +43,11 @@ jobs:
       - telemetry-setup
       - third-party-integration-tests-cudf-pandas
     secrets: inherit
+<<<<<<< HEAD
     uses: rapidsai/shared-workflows/.github/workflows/[email protected]
+=======
+    uses: rapidsai/shared-workflows/.github/workflows/[email protected]
+>>>>>>> branch-25.06
     if: always()
     with:
       needs: ${{ toJSON(needs) }}

ci/cudf_pandas_scripts/fetch_pandas_versions.py

@@ -3,25 +3,55 @@
 import argparse
 import json
 import ssl
+import sys
 import urllib.request
 
 import certifi
 from packaging.specifiers import SpecifierSet
 from packaging.version import Version
 
 
+def has_python_version(version, data):
+    for d in data:
+        try:
+            if Version(d["python_version"][2:]) == version:
+                return True
+        except Exception:
+            pass
+    return False
+
+
 def get_pandas_versions(pandas_range):
     url = "https://pypi.org/pypi/pandas/json"
     ssl_context = ssl.create_default_context(cafile=certifi.where())
-    with urllib.request.urlopen(url, context=ssl_context) as response:
+    python_version = Version(
+        str(sys.version_info.major) + str(sys.version_info.minor)
+    )
+    # Set a timeout for the request to avoid hanging
+    timeout = 10  # seconds
+
+    # Try to fetch pandas versions from PyPI
+    with urllib.request.urlopen(
+        url, timeout=timeout, context=ssl_context
+    ) as response:
         data = json.loads(response.read())
+
+    # Extract and filter versions
     versions = [Version(v) for v in data["releases"]]
+
     specifier = SpecifierSet(pandas_range.lstrip("pandas"))
-    matching_versions = [v for v in versions if v in specifier]
+    matching_versions = [
+        v
+        for v in versions
+        if v in specifier
+        and has_python_version(python_version, data["releases"][str(v)])
+    ]
+
     matching_minors = sorted(
         set(".".join((str(v.major), str(v.minor))) for v in matching_versions),
         key=Version,
     )
+
     return matching_minors
 
 

conda/environments/all_cuda-118_arch-aarch64.yaml

@@ -54,7 +54,7 @@ dependencies:
 - nbsphinx
 - ninja
 - notebook
-- numba-cuda>=0.10.1,<0.11.0a0
+- numba-cuda>=0.11.0,<0.12.0a0
 - numba>=0.59.1,<0.62.0a0
 - numpy>=1.23,<3.0a0
 - numpydoc

conda/environments/all_cuda-118_arch-x86_64.yaml

@@ -56,7 +56,7 @@ dependencies:
 - nbsphinx
 - ninja
 - notebook
-- numba-cuda>=0.10.1,<0.11.0a0
+- numba-cuda>=0.11.0,<0.12.0a0
 - numba>=0.59.1,<0.62.0a0
 - numpy>=1.23,<3.0a0
 - numpydoc

conda/environments/all_cuda-128_arch-aarch64.yaml

@@ -54,7 +54,7 @@ dependencies:
 - nbsphinx
 - ninja
 - notebook
-- numba-cuda>=0.10.1,<0.11.0a0
+- numba-cuda>=0.11.0,<0.12.0a0
 - numba>=0.59.1,<0.62.0a0
 - numpy>=1.23,<3.0a0
 - numpydoc

conda/environments/all_cuda-128_arch-x86_64.yaml

@@ -55,7 +55,7 @@ dependencies:
 - nbsphinx
 - ninja
 - notebook
-- numba-cuda>=0.10.1,<0.11.0a0
+- numba-cuda>=0.11.0,<0.12.0a0
 - numba>=0.59.1,<0.62.0a0
 - numpy>=1.23,<3.0a0
 - numpydoc

conda/recipes/cudf/recipe.yaml

@@ -73,7 +73,7 @@ requirements:
     - typing_extensions >=4.0.0
     - pandas >=2.0,<2.2.4dev0
     - cupy >=12.0.0
-    - numba-cuda >=0.10.1,<0.11.0a0
+    - numba-cuda >=0.11.0,<0.12.0a0
     - numba >=0.59.1,<0.62.0a0
     - numpy >=1.23,<3.0a0
     - pyarrow>=14.0.0,<20.0.0a0

cpp/include/cudf/detail/join/distinct_hash_join.cuh

@@ -49,7 +49,7 @@ struct always_not_equal {
 };
 
 /**
- * @brief An comparator adapter wrapping the two table comparator
+ * @brief A comparator adapter wrapping the two table comparator
  */
 template <typename Equal>
 struct comparator_adapter {
@@ -67,6 +67,25 @@ struct comparator_adapter {
   Equal _d_equal;
 };
 
+/**
+ * @brief A comparator adapter wrapping the two table comparator
+ */
+template <typename Equal>
+struct primitive_comparator_adapter {
+  primitive_comparator_adapter(Equal const& d_equal) : _d_equal{d_equal} {}
+
+  __device__ constexpr auto operator()(
+    cuco::pair<hash_value_type, lhs_index_type> const& lhs,
+    cuco::pair<hash_value_type, rhs_index_type> const& rhs) const noexcept
+  {
+    if (lhs.first != rhs.first) { return false; }
+    return _d_equal(static_cast<size_type>(lhs.second), static_cast<size_type>(rhs.second));
+  }
+
+ private:
+  Equal _d_equal;
+};
+
 /**
  * @brief Distinct hash join that builds hash table in creation and probes results in subsequent
  * `*_join` member functions.

cpp/include/cudf/table/experimental/row_operators.cuh

@@ -56,6 +56,13 @@
 
 namespace CUDF_EXPORT cudf {
 
+namespace row::primitive {
+class row_equality_comparator;  // Forward declaration
+
+template <template <typename> class Hash>
+class row_hasher;  // Forward declaration
+}  // namespace row::primitive
+
 namespace experimental {
 
 /**
@@ -1577,6 +1584,11 @@ struct preprocessed_table {
   friend class self_comparator;       ///< Allow self_comparator to access private members
   friend class two_table_comparator;  ///< Allow two_table_comparator to access private members
   friend class hash::row_hasher;      ///< Allow row_hasher to access private members
+  /// Allow primitive equality comparator to access private members
+  friend class ::cudf::row::primitive::row_equality_comparator;
+
+  template <template <typename> class Hash>
+  friend class ::cudf::row::primitive::row_hasher;
 
   using table_device_view_owner =
     std::invoke_result_t<decltype(table_device_view::create), table_view, rmm::cuda_stream_view>;