GH-37710: [C++][Integration] Add C++ Utf8View implementation

apache · Sep 20, 2023 · 660008f · 660008f
1 parent b01b801
commit 660008f
Show file tree

Hide file tree

Showing 70 changed files with 2,571 additions and 345 deletions.
diff --git a/cpp/src/arrow/array/array_base.cc b/cpp/src/arrow/array/array_base.cc
@@ -87,6 +87,8 @@ struct ScalarFromArraySlotImpl {
     return Finish(a.GetString(index_));
   }
 
+  Status Visit(const BinaryViewArray& a) { return Finish(a.GetString(index_)); }
+
   Status Visit(const FixedSizeBinaryArray& a) { return Finish(a.GetString(index_)); }
 
   Status Visit(const DayTimeIntervalArray& a) { return Finish(a.Value(index_)); }

diff --git a/cpp/src/arrow/array/array_binary.cc b/cpp/src/arrow/array/array_binary.cc
@@ -24,6 +24,7 @@
 #include "arrow/array/validate.h"
 #include "arrow/type.h"
 #include "arrow/type_traits.h"
+#include "arrow/util/binary_view_util.h"
 #include "arrow/util/checked_cast.h"
 #include "arrow/util/logging.h"
 
@@ -89,6 +90,37 @@ LargeStringArray::LargeStringArray(int64_t length,
 
 Status LargeStringArray::ValidateUTF8() const { return internal::ValidateUTF8(*data_); }
 
+BinaryViewArray::BinaryViewArray(const std::shared_ptr<ArrayData>& data) {
+  ARROW_CHECK_EQ(data->type->id(), Type::BINARY_VIEW);
+  SetData(data);
+}
+
+BinaryViewArray::BinaryViewArray(int64_t length, std::shared_ptr<Buffer> headers,
+                                 BufferVector char_buffers,
+                                 std::shared_ptr<Buffer> null_bitmap, int64_t null_count,
+                                 int64_t offset)
+    : PrimitiveArray(binary_view(), length, std::move(headers), std::move(null_bitmap),
+                     null_count, offset) {
+  data_->buffers.resize(char_buffers.size() + 2);
+  std::move(char_buffers.begin(), char_buffers.end(), data_->buffers.begin() + 2);
+}
+
+std::string_view BinaryViewArray::GetView(int64_t i) const {
+  const auto& s = raw_values()[i];
+  if (raw_pointers_) {
+    return util::FromRawPointerBinaryView(s);
+  }
+  const std::shared_ptr<Buffer>* data_buffers = data_->buffers.data() + 2;
+  return util::FromIndexOffsetBinaryView(s, data_buffers);
+}
+
+StringViewArray::StringViewArray(const std::shared_ptr<ArrayData>& data) {
+  ARROW_CHECK_EQ(data->type->id(), Type::STRING_VIEW);
+  SetData(data);
+}
+
+Status StringViewArray::ValidateUTF8() const { return internal::ValidateUTF8(*data_); }
+
 FixedSizeBinaryArray::FixedSizeBinaryArray(const std::shared_ptr<ArrayData>& data) {
   SetData(data);
 }

diff --git a/cpp/src/arrow/array/array_binary.h b/cpp/src/arrow/array/array_binary.h
@@ -22,6 +22,7 @@
 
 #include <cstdint>
 #include <memory>
+#include <optional>
 #include <string>
 #include <string_view>
 #include <vector>
@@ -217,6 +218,75 @@ class ARROW_EXPORT LargeStringArray : public LargeBinaryArray {
   Status ValidateUTF8() const;
 };
 
+// ----------------------------------------------------------------------
+// BinaryView and StringView
+
+/// Concrete Array class for variable-size binary view data using the
+/// BinaryViewType::c_type struct to reference in-line or out-of-line string values
+class ARROW_EXPORT BinaryViewArray : public PrimitiveArray {
+ public:
+  using TypeClass = BinaryViewType;
+  using IteratorType = stl::ArrayIterator<BinaryViewArray>;
+  using c_type = BinaryViewType::c_type;
+
+  explicit BinaryViewArray(const std::shared_ptr<ArrayData>& data);
+
+  BinaryViewArray(int64_t length, std::shared_ptr<Buffer> headers,
+                  BufferVector char_buffers,
+                  std::shared_ptr<Buffer> null_bitmap = NULLPTR,
+                  int64_t null_count = kUnknownNullCount, int64_t offset = 0);
+
+  const c_type* raw_values() const {
+    return reinterpret_cast<const c_type*>(raw_values_) + data_->offset;
+  }
+
+  // For API compatibility with BinaryArray etc.
+  std::string_view GetView(int64_t i) const;
+  std::string GetString(int64_t i) const { return std::string{GetView(i)}; }
+
+  std::optional<std::string_view> operator[](int64_t i) const {
+    return *IteratorType(*this, i);
+  }
+
+  IteratorType begin() const { return IteratorType(*this); }
+  IteratorType end() const { return IteratorType(*this, length()); }
+
+  bool has_raw_pointers() const { return raw_pointers_; }
+
+ protected:
+  using PrimitiveArray::PrimitiveArray;
+
+  void SetData(const std::shared_ptr<ArrayData>& data) {
+    PrimitiveArray::SetData(data);
+    raw_pointers_ =
+        internal::checked_cast<const BinaryViewType&>(*type()).has_raw_pointers();
+  }
+
+  bool raw_pointers_ = false;
+};
+
+/// Concrete Array class for variable-size string view (utf-8) data using
+/// BinaryViewType::c_type to reference in-line or out-of-line string values
+class ARROW_EXPORT StringViewArray : public BinaryViewArray {
+ public:
+  using TypeClass = StringViewType;
+
+  explicit StringViewArray(const std::shared_ptr<ArrayData>& data);
+
+  StringViewArray(int64_t length, std::shared_ptr<Buffer> data, BufferVector char_buffers,
+                  std::shared_ptr<Buffer> null_bitmap = NULLPTR,
+                  int64_t null_count = kUnknownNullCount, int64_t offset = 0)
+      : BinaryViewArray(length, std::move(data), std::move(char_buffers),
+                        std::move(null_bitmap), null_count, offset) {
+    data_->type = utf8_view();
+  }
+
+  /// \brief Validate that this array contains only valid UTF8 entries
+  ///
+  /// This check is also implied by ValidateFull()
+  Status ValidateUTF8() const;
+};
+
 // ----------------------------------------------------------------------
 // Fixed width binary