From 3f9a883897e996aaf04b220f73d3b532528fa241 Mon Sep 17 00:00:00 2001
From: Sutou Kouhei <kou@clear-code.com>
Date: Tue, 16 Dec 2025 14:20:28 +0900
Subject: [PATCH] GH-48551: [Ruby] Add support for reading large UTF-8 array

---
 .../lib/arrow-format/array.rb                 | 11 ++++++++++
 .../lib/arrow-format/file-reader.rb           |  4 ++++
 .../red-arrow-format/lib/arrow-format/type.rb | 20 +++++++++++++++++++
 .../red-arrow-format/test/test-file-reader.rb | 11 ++++++++++
 4 files changed, 46 insertions(+)

diff --git a/ruby/red-arrow-format/lib/arrow-format/array.rb b/ruby/red-arrow-format/lib/arrow-format/array.rb
index 23969bc24bd..edf8326f3d4 100644
--- a/ruby/red-arrow-format/lib/arrow-format/array.rb
+++ b/ruby/red-arrow-format/lib/arrow-format/array.rb
@@ -235,6 +235,17 @@ def encoding
     end
   end
 
+  class LargeUTF8Array < VariableSizeBinaryLayoutArray
+    private
+    def buffer_type
+      :s64 # TODO: big endian support
+    end
+
+    def encoding
+      Encoding::UTF_8
+    end
+  end
+
   class FixedSizeBinaryArray < Array
     def initialize(type, size, validity_buffer, values_buffer)
       super(type, size, validity_buffer)
diff --git a/ruby/red-arrow-format/lib/arrow-format/file-reader.rb b/ruby/red-arrow-format/lib/arrow-format/file-reader.rb
index 8b149cd1756..f08e3e448e3 100644
--- a/ruby/red-arrow-format/lib/arrow-format/file-reader.rb
+++ b/ruby/red-arrow-format/lib/arrow-format/file-reader.rb
@@ -26,11 +26,13 @@
 require_relative "org/apache/arrow/flatbuf/bool"
 require_relative "org/apache/arrow/flatbuf/date"
 require_relative "org/apache/arrow/flatbuf/date_unit"
+require_relative "org/apache/arrow/flatbuf/fixed_size_binary"
 require_relative "org/apache/arrow/flatbuf/floating_point"
 require_relative "org/apache/arrow/flatbuf/footer"
 require_relative "org/apache/arrow/flatbuf/int"
 require_relative "org/apache/arrow/flatbuf/large_binary"
 require_relative "org/apache/arrow/flatbuf/large_list"
+require_relative "org/apache/arrow/flatbuf/large_utf8"
 require_relative "org/apache/arrow/flatbuf/list"
 require_relative "org/apache/arrow/flatbuf/map"
 require_relative "org/apache/arrow/flatbuf/message"
@@ -232,6 +234,8 @@ def read_field(fb_field)
         type = LargeBinaryType.singleton
       when Org::Apache::Arrow::Flatbuf::Utf8
         type = UTF8Type.singleton
+      when Org::Apache::Arrow::Flatbuf::LargeUtf8
+        type = LargeUTF8Type.singleton
       when Org::Apache::Arrow::Flatbuf::FixedSizeBinary
         type = FixedSizeBinaryType.new(fb_type.byte_width)
       end
diff --git a/ruby/red-arrow-format/lib/arrow-format/type.rb b/ruby/red-arrow-format/lib/arrow-format/type.rb
index 055c0890c4b..a66fe08f25a 100644
--- a/ruby/red-arrow-format/lib/arrow-format/type.rb
+++ b/ruby/red-arrow-format/lib/arrow-format/type.rb
@@ -353,6 +353,26 @@ def build_array(size, validity_buffer, offsets_buffer, values_buffer)
     end
   end
 
+  class LargeUTF8Type < VariableSizeBinaryType
+    class << self
+      def singleton
+        @singleton ||= new
+      end
+    end
+
+    def initialize
+      super("LargeUTF8")
+    end
+
+    def build_array(size, validity_buffer, offsets_buffer, values_buffer)
+      LargeUTF8Array.new(self,
+                         size,
+                         validity_buffer,
+                         offsets_buffer,
+                         values_buffer)
+    end
+  end
+
   class FixedSizeBinaryType < Type
     attr_reader :byte_width
     def initialize(byte_width)
diff --git a/ruby/red-arrow-format/test/test-file-reader.rb b/ruby/red-arrow-format/test/test-file-reader.rb
index b31e8940458..c66948518bb 100644
--- a/ruby/red-arrow-format/test/test-file-reader.rb
+++ b/ruby/red-arrow-format/test/test-file-reader.rb
@@ -358,6 +358,17 @@ def test_read
     end
   end
 
+  sub_test_case("LargeUTF8") do
+    def build_array
+      Arrow::LargeStringArray.new(["Hello", nil, "World"])
+    end
+
+    def test_read
+      assert_equal([{"value" => ["Hello", nil, "World"]}],
+                   read)
+    end
+  end
+
   sub_test_case("FixedSizeBinary") do
     def build_array
       data_type = Arrow::FixedSizeBinaryDataType.new(4)