GH-48945: [Ruby] Add support for writing large binary array

Author: kou

ruby/red-arrow-format/lib/arrow-format/type.rb

@@ -505,6 +505,10 @@ def build_array(size, validity_buffer, offsets_buffer, values_buffer)
                            offsets_buffer,
                            values_buffer)
     end
+
+    def to_flatbuffers
+      FB::LargeBinary::Data.new
+    end
   end
 
   class UTF8Type < VariableSizeBinaryType

ruby/red-arrow-format/test/test-writer.rb

@@ -40,6 +40,8 @@ def convert_type(red_arrow_type)
       ArrowFormat::UInt64Type.singleton
     when Arrow::BinaryDataType
       ArrowFormat::BinaryType.singleton
+    when Arrow::LargeBinaryDataType
+      ArrowFormat::LargeBinaryType.singleton
     when Arrow::StringDataType
       ArrowFormat::UTF8Type.singleton
     else
@@ -203,6 +205,17 @@ def test_write
           end
         end
 
+        sub_test_case("LargeBinary") do
+          def build_array
+            Arrow::LargeBinaryArray.new(["Hello".b, nil, "World".b])
+          end
+
+          def test_write
+            assert_equal(["Hello".b, nil, "World".b],
+                         @values)
+          end
+        end
+
         sub_test_case("String") do
           def build_array
             Arrow::StringArray.new(["Hello", nil, "World"])

ruby/red-arrow/ext/arrow/converters.hpp

@@ -153,7 +153,15 @@ namespace red_arrow {
                          const int64_t i) {
       int32_t length;
       const auto value = array.GetValue(i, &length);
-      // TODO: encoding support
+      return rb_enc_str_new(reinterpret_cast<const char*>(value),
+                            length,
+                            rb_ascii8bit_encoding());
+    }
+
+    inline VALUE convert(const arrow::LargeBinaryArray& array,
+                         const int64_t i) {
+      int64_t length;
+      const auto value = array.GetValue(i, &length);
       return rb_enc_str_new(reinterpret_cast<const char*>(value),
                             length,
                             rb_ascii8bit_encoding());

ruby/red-arrow/ext/arrow/raw-records.cpp

@@ -88,6 +88,7 @@ namespace red_arrow {
       VISIT(Float)
       VISIT(Double)
       VISIT(Binary)
+      VISIT(LargeBinary)
       VISIT(String)
       VISIT(FixedSizeBinary)
       VISIT(Date32)
@@ -224,6 +225,7 @@ namespace red_arrow {
       VISIT(Float)
       VISIT(Double)
       VISIT(Binary)
+      VISIT(LargeBinary)
       VISIT(String)
       VISIT(FixedSizeBinary)
       VISIT(Date32)

ruby/red-arrow/ext/arrow/values.cpp

@@ -69,6 +69,7 @@ namespace red_arrow {
       VISIT(Float)
       VISIT(Double)
       VISIT(Binary)
+      VISIT(LargeBinary)
       VISIT(String)
       VISIT(FixedSizeBinary)
       VISIT(Date32)

ruby/red-arrow/test/raw-records/test-basic-arrays.rb

@@ -157,6 +157,16 @@ def test_binary
     assert_equal(records, actual_records(target))
   end
 
+  def test_large_binary
+    records = [
+      ["\x00".b],
+      [nil],
+      ["\xff".b],
+    ]
+    target = build({column: :large_binary}, records)
+    assert_equal(records, actual_records(target))
+  end
+
   def test_string
     records = [
       ["Ruby"],

ruby/red-arrow/test/values/test-basic-arrays.rb

@@ -147,7 +147,17 @@ def test_binary
     assert_equal(values, target.values)
   end
 
-  def test_tring
+  def test_large_binary
+    values = [
+      "\x00".b,
+      nil,
+      "\xff".b,
+    ]
+    target = build(Arrow::LargeBinaryArray.new(values))
+    assert_equal(values, target.values)
+  end
+
+  def test_string
     values = [
       "Ruby",
       nil,