This is an automated email from the ASF dual-hosted git repository.
zanmato pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/arrow.git
The following commit(s) were added to refs/heads/main by this push:
new 37eb3b5010 GH-45005 [C++] Support for fixed-size list in conversion of
range tuple (#45008)
37eb3b5010 is described below
commit 37eb3b50101aa5e32b5122e61593a1a592a22274
Author: mroz45 <[email protected]>
AuthorDate: Wed Dec 18 11:14:03 2024 +0100
GH-45005 [C++] Support for fixed-size list in conversion of range tuple
(#45008)
### What changes are included in this PR?
This PR introduces support for fixed-size lists in Apache Arrow's STL
conversion. Added specializations for arrow::stl::ConversionTraits<T> and
arrow::CTypeTraits<T> to handle std::array as a fixed-size list and also Unit
Test.
### Are these changes tested?
Yes, new type is tested.
### Are there any user-facing changes?
* GitHub Issue: #45005
Lead-authored-by: kamilt <[email protected]>
Co-authored-by: Rossi Sun <[email protected]>
Signed-off-by: Rossi Sun <[email protected]>
---
cpp/src/arrow/stl.h | 29 +++++++++++++++++++++++++++++
cpp/src/arrow/stl_test.cc | 40 ++++++++++++++++++++++++++++++++++++++++
cpp/src/arrow/type_traits.h | 10 ++++++++++
3 files changed, 79 insertions(+)
diff --git a/cpp/src/arrow/stl.h b/cpp/src/arrow/stl.h
index b542ee5c34..ae5462c661 100644
--- a/cpp/src/arrow/stl.h
+++ b/cpp/src/arrow/stl.h
@@ -187,6 +187,35 @@ struct ConversionTraits<std::vector<ValueCType>>
}
};
+template <class ValueCType, std::size_t N>
+struct ConversionTraits<std::array<ValueCType, N>>
+ : public CTypeTraits<std::array<ValueCType, N>> {
+ static arrow::Status AppendRow(FixedSizeListBuilder& builder,
+ const std::array<ValueCType, N>& values) {
+ auto vb =
+ ::arrow::internal::checked_cast<typename
CTypeTraits<ValueCType>::BuilderType*>(
+ builder.value_builder());
+ ARROW_RETURN_NOT_OK(builder.Append());
+ return vb->AppendValues(values.data(), N);
+ }
+
+ static std::array<ValueCType, N> GetEntry(const ::arrow::FixedSizeListArray&
array,
+ size_t j) {
+ using ElementArrayType = typename TypeTraits<
+ typename stl::ConversionTraits<ValueCType>::ArrowType>::ArrayType;
+
+ const ElementArrayType& value_array =
+ ::arrow::internal::checked_cast<const
ElementArrayType&>(*array.values());
+
+ std::array<ValueCType, N> arr;
+ for (size_t i = 0; i < N; i++) {
+ arr[i] = stl::ConversionTraits<ValueCType>::GetEntry(value_array,
+
array.value_offset(j) + i);
+ }
+ return arr;
+ }
+};
+
template <typename Optional>
struct ConversionTraits<Optional, enable_if_optional_like<Optional>>
: public CTypeTraits<typename
std::decay<decltype(*std::declval<Optional>())>::type> {
diff --git a/cpp/src/arrow/stl_test.cc b/cpp/src/arrow/stl_test.cc
index 48e6f8014c..ce5adf0c0e 100644
--- a/cpp/src/arrow/stl_test.cc
+++ b/cpp/src/arrow/stl_test.cc
@@ -245,6 +245,26 @@ TEST(TestTableFromTupleVector, ListType) {
ASSERT_TRUE(expected_table->Equals(*table));
}
+TEST(TestTableFromTupleVector, FixedSizeListType) {
+ using tuple_type = std::tuple<std::array<int64_t, 4>>;
+
+ auto expected_schema = std::make_shared<Schema>(
+ FieldVector{field("column1", fixed_size_list(int64(), 4), false)});
+ std::shared_ptr<Array> expected_array =
+ ArrayFromJSON(fixed_size_list(int64(), 4), "[[1, 1, 2, 34], [2, -4, 1,
1]]");
+ std::shared_ptr<Table> expected_table = Table::Make(expected_schema,
{expected_array});
+
+ std::vector<tuple_type> rows{tuple_type(std::array<int64_t, 4>{1, 1, 2, 34}),
+ tuple_type(std::array<int64_t, 4>{2, -4, 1,
1})};
+ std::vector<std::string> names{"column1"};
+
+ std::shared_ptr<Table> table;
+ ASSERT_OK(TableFromTupleRange(default_memory_pool(), rows, names, &table));
+ ASSERT_OK(table->ValidateFull());
+
+ AssertTablesEqual(*expected_table, *table);
+}
+
TEST(TestTableFromTupleVector, ReferenceTuple) {
std::vector<std::string> names{"column1", "column2", "column3", "column4",
"column5",
"column6", "column7", "column8", "column9",
"column10"};
@@ -468,6 +488,26 @@ TEST(TestTupleVectorFromTable, ListType) {
ASSERT_EQ(rows, expected_rows);
}
+TEST(TestTupleVectorFromTable, FixedSizeListType) {
+ using tuple_type = std::tuple<std::array<int64_t, 4>>;
+
+ compute::ExecContext ctx;
+ compute::CastOptions cast_options;
+ auto expected_schema = std::make_shared<Schema>(
+ FieldVector{field("column1", fixed_size_list(int64(), 4), false)});
+ std::shared_ptr<Array> expected_array =
+ ArrayFromJSON(fixed_size_list(int64(), 4), "[[1, 1, 2, 34], [2, -4, 1,
1]]");
+ std::shared_ptr<Table> table = Table::Make(expected_schema,
{expected_array});
+ ASSERT_OK(table->ValidateFull());
+
+ std::vector<tuple_type> expected_rows{tuple_type(std::array<int64_t, 4>{1,
1, 2, 34}),
+ tuple_type(std::array<int64_t, 4>{2,
-4, 1, 1})};
+
+ std::vector<tuple_type> rows(2);
+ ASSERT_OK(TupleRangeFromTable(*table, cast_options, &ctx, &rows));
+ ASSERT_EQ(rows, expected_rows);
+}
+
TEST(TestTupleVectorFromTable, CastingNeeded) {
using tuple_type = std::tuple<std::vector<int64_t>>;
diff --git a/cpp/src/arrow/type_traits.h b/cpp/src/arrow/type_traits.h
index 6da05bd8f1..92009c8560 100644
--- a/cpp/src/arrow/type_traits.h
+++ b/cpp/src/arrow/type_traits.h
@@ -540,6 +540,16 @@ struct CTypeTraits<std::vector<CType>> : public
TypeTraits<ListType> {
}
};
+/// \addtogroup c-type-traits
+template <typename CType, std::size_t N>
+struct CTypeTraits<std::array<CType, N>> : public
TypeTraits<FixedSizeListType> {
+ using ArrowType = FixedSizeListType;
+
+ static auto type_singleton() {
+ return fixed_size_list(CTypeTraits<CType>::type_singleton(), N);
+ }
+};
+
/// \addtogroup type-traits
/// @{
template <>
