Add binary_type to ArrowReaderProperties (#569)

Author: haianhng31

cpp/Enums.cpp

@@ -102,5 +102,9 @@ namespace
 		static_assert((int) ::arrow::StatusCode::OutOfMemory == 1);
 		static_assert((int) ::arrow::StatusCode::IOError == 5);
 		static_assert((int) ::arrow::StatusCode::UnknownError == 9);
+
+		static_assert((int) ::arrow::Type::type::BINARY == 14);
+		static_assert((int) ::arrow::Type::type::LARGE_BINARY == 35);
+		static_assert((int) ::arrow::Type::type::BINARY_VIEW == 40);
 	}
 }

cpp/arrow/ArrowReaderProperties.cpp

@@ -66,4 +66,14 @@ extern "C"
   {
     TRYCATCH(properties->set_coerce_int96_timestamp_unit(unit);)
   }
+
+  PARQUETSHARP_EXPORT ExceptionInfo* ArrowReaderProperties_BinaryType(ArrowReaderProperties* properties, ::arrow::Type::type* value)
+  {
+    TRYCATCH(*value = properties->binary_type();)
+  }
+
+  PARQUETSHARP_EXPORT ExceptionInfo* ArrowReaderProperties_SetBinaryType(ArrowReaderProperties* properties, ::arrow::Type::type value)
+  {
+    TRYCATCH(properties->set_binary_type(value);)
+  }
 }

csharp.test/Arrow/TestArrowReaderProperties.cs

@@ -28,12 +28,14 @@ public void TestSetProperties()
             properties.SetReadDictionary(0, true);
             properties.PreBuffer = false;
             properties.CoerceInt96TimestampUnit = Apache.Arrow.Types.TimeUnit.Microsecond;
+            properties.BinaryType = Apache.Arrow.Types.ArrowTypeId.LargeBinary;
 
             Assert.That(properties.UseThreads, Is.True);
             Assert.That(properties.BatchSize, Is.EqualTo(789));
             Assert.That(properties.GetReadDictionary(0), Is.True);
             Assert.That(properties.PreBuffer, Is.False);
             Assert.That(properties.CoerceInt96TimestampUnit, Is.EqualTo(Apache.Arrow.Types.TimeUnit.Microsecond));
+            Assert.That(properties.BinaryType, Is.EqualTo(Apache.Arrow.Types.ArrowTypeId.LargeBinary));
         }
     }
 }

csharp/Arrow/ArrowReaderProperties.cs

@@ -105,6 +105,34 @@ public Apache.Arrow.Types.TimeUnit CoerceInt96TimestampUnit
             }
         }
 
+        /// <summary>
+        /// The Arrow binary type to read BYTE_ARRAY columns as.
+        /// 
+        /// Allowed values are ArrowTypeId.Binary, ArrowTypeId.LargeBinary and ArrowTypeId.BinaryView.
+        /// Default is ArrowTypeId.Binary.
+        ///
+        /// If a BYTE_ARRAY column has the STRING logical type, it is read as the
+        /// Arrow string type corresponding to the configured binary type (for example
+        /// Type::LARGE_STRING if the configured binary type is Type::LARGE_BINARY).
+        ///
+        /// However, if a serialized Arrow schema is found in the Parquet metadata,
+        /// this setting is ignored and the Arrow schema takes precedence
+        /// </summary>
+        public Apache.Arrow.Types.ArrowTypeId BinaryType
+        {
+            get
+            {
+                ParquetSharp.CppTypeId value = ExceptionInfo.Return<ParquetSharp.CppTypeId>(Handle, ArrowReaderProperties_BinaryType);
+                return value.toPublicEnum();
+            }
+            set
+            {
+                ParquetSharp.CppTypeId cppValue = value.toCppEnum();
+                ExceptionInfo.Check(ArrowReaderProperties_SetBinaryType(Handle.IntPtr, cppValue));
+                GC.KeepAlive(Handle);
+            }
+        }
+
         [DllImport(ParquetDll.Name)]
         private static extern IntPtr ArrowReaderProperties_GetDefault(out IntPtr readerProperties);
 
@@ -141,6 +169,12 @@ public Apache.Arrow.Types.TimeUnit CoerceInt96TimestampUnit
         [DllImport(ParquetDll.Name)]
         private static extern IntPtr ArrowReaderProperties_SetCoerceInt96TimestampUnit(IntPtr readerProperties, Apache.Arrow.Types.TimeUnit unit);
 
+        [DllImport(ParquetDll.Name)]
+        private static extern IntPtr ArrowReaderProperties_BinaryType(IntPtr readerProperties, out ParquetSharp.CppTypeId value);
+
+        [DllImport(ParquetDll.Name)]
+        private static extern IntPtr ArrowReaderProperties_SetBinaryType(IntPtr readerProperties, ParquetSharp.CppTypeId value);
+
         internal readonly ParquetHandle Handle;
     }
 }

csharp/ArrowTypeIdExtensions.cs

@@ -0,0 +1,16 @@
+using System;
+using System.Runtime.InteropServices;
+
+namespace ParquetSharp
+{
+    internal static class ArrowTypeIdExtensions
+    {
+        public static ParquetSharp.CppTypeId toCppEnum(this Apache.Arrow.Types.ArrowTypeId arrowTypeId) => arrowTypeId switch
+        {
+            Apache.Arrow.Types.ArrowTypeId.Binary => ParquetSharp.CppTypeId.Binary,
+            Apache.Arrow.Types.ArrowTypeId.LargeBinary => ParquetSharp.CppTypeId.LargeBinary,
+            Apache.Arrow.Types.ArrowTypeId.BinaryView => ParquetSharp.CppTypeId.BinaryView,
+            _ => throw new ArgumentOutOfRangeException(nameof(arrowTypeId), arrowTypeId, null)
+        };
+    }
+}

csharp/CppTypeId.cs

@@ -0,0 +1,23 @@
+using System;
+using System.Runtime.InteropServices;
+
+namespace ParquetSharp
+{
+    internal enum CppTypeId
+    {
+        Binary = 14,
+        LargeBinary = 35,
+        BinaryView = 40,
+    }
+
+    internal static class CppTypeIdExtensions
+    {
+        public static Apache.Arrow.Types.ArrowTypeId toPublicEnum(this CppTypeId binaryType) => binaryType switch
+        {
+            CppTypeId.Binary => Apache.Arrow.Types.ArrowTypeId.Binary,
+            CppTypeId.LargeBinary => Apache.Arrow.Types.ArrowTypeId.LargeBinary,
+            CppTypeId.BinaryView => Apache.Arrow.Types.ArrowTypeId.BinaryView,
+            _ => throw new ArgumentOutOfRangeException(nameof(binaryType), binaryType, null)
+        };
+    }
+}

csharp/PublicAPI.Unshipped.txt

@@ -1,7 +1,9 @@
 #nullable enable
+ParquetSharp.Arrow.ArrowReaderProperties.BinaryType.get -> Apache.Arrow.Types.ArrowTypeId
+ParquetSharp.Arrow.ArrowReaderProperties.BinaryType.set -> void
 ParquetSharp.ReaderProperties.ThriftStringSizeLimit.get -> int
 ParquetSharp.ReaderProperties.SetThriftStringSizeLimit(int size) -> void
 ParquetSharp.ReaderProperties.ThriftContainerSizeLimit.get -> int
 ParquetSharp.ReaderProperties.SetThriftContainerSizeLimit(int size) -> void
 ParquetSharp.ReaderProperties.FooterReadSize.get -> long
-ParquetSharp.ReaderProperties.SetFooterReadSize(long size) -> void
+ParquetSharp.ReaderProperties.SetFooterReadSize(long size) -> void