Make VariantLogicalType singleton

Author: voonhous

hudi-common/src/main/java/org/apache/hudi/common/schema/HoodieSchema.java

@@ -128,7 +128,7 @@ public static HoodieSchema fromAvroSchema(Schema avroSchema) {
       } else if (logicalType instanceof LogicalTypes.TimestampMillis || logicalType instanceof LogicalTypes.TimestampMicros
           || logicalType instanceof LogicalTypes.LocalTimestampMillis || logicalType instanceof LogicalTypes.LocalTimestampMicros) {
         return new HoodieSchema.Timestamp(avroSchema);
-      } else if (logicalType instanceof VariantLogicalType) {
+      } else if (logicalType == VariantLogicalType.variant()) {
         return new HoodieSchema.Variant(avroSchema);
       }
     }
@@ -531,7 +531,7 @@ public static HoodieSchema.Variant createVariant(String name, String namespace,
     recordSchema.setFields(avroFields);
 
     // Add Variant logical type
-    new VariantLogicalType().addToSchema(recordSchema);
+    VariantLogicalType.variant().addToSchema(recordSchema);
 
     return new HoodieSchema.Variant(recordSchema);
   }
@@ -594,7 +594,7 @@ public static HoodieSchema.Variant createVariantShredded(String name, String nam
     recordSchema.setFields(avroFields);
 
     // Add Variant logical type
-    new VariantLogicalType().addToSchema(recordSchema);
+    VariantLogicalType.variant().addToSchema(recordSchema);
 
     return new HoodieSchema.Variant(recordSchema);
   }
@@ -1479,15 +1479,28 @@ public enum TimePrecision {
   /**
    * Custom Avro LogicalType for Variant.
    * This logical type is applied to RECORD schemas that represent Variant types.
+   *
+   * <p>This is a singleton type - use {@link #variant()} to get the instance.</p>
    */
   public static class VariantLogicalType extends LogicalType {
 
     private static final String VARIANT_LOGICAL_TYPE_NAME = "variant";
+    // Eager initialization of singleton
+    private static final VariantLogicalType INSTANCE = new VariantLogicalType();
 
-    public VariantLogicalType() {
+    private VariantLogicalType() {
       super(VariantLogicalType.VARIANT_LOGICAL_TYPE_NAME);
     }
 
+    /**
+     * Returns the singleton instance of VariantLogicalType.
+     *
+     * @return the Variant logical type instance
+     */
+    public static VariantLogicalType variant() {
+      return INSTANCE;
+    }
+
     @Override
     public void validate(Schema schema) {
       super.validate(schema);
@@ -1503,7 +1516,7 @@ public void validate(Schema schema) {
   private static class VariantLogicalTypeFactory implements LogicalTypes.LogicalTypeFactory {
     @Override
     public LogicalType fromSchema(Schema schema) {
-      return new VariantLogicalType();
+      return VariantLogicalType.variant();
     }
 
     @Override

hudi-common/src/test/java/org/apache/hudi/common/schema/TestHoodieSchema.java

@@ -22,6 +22,7 @@
 import org.apache.hudi.exception.HoodieAvroSchemaException;
 
 import org.apache.avro.JsonProperties;
+import org.apache.avro.LogicalType;
 import org.apache.avro.LogicalTypes;
 import org.apache.avro.Schema;
 import org.junit.jupiter.api.Test;
@@ -1529,4 +1530,40 @@ public void testVariantCustomNameAndNamespace() {
     assertTrue(HoodieSchema.Variant.isVariantSchema(customVariant.getAvroSchema()));
     assertEquals(HoodieSchemaType.VARIANT, customVariant.getType());
   }
+
+  @Test
+  public void testVariantLogicalTypeSingleton() {
+    // Test that VariantLogicalType is a true singleton
+    HoodieSchema.VariantLogicalType instance1 = HoodieSchema.VariantLogicalType.variant();
+    HoodieSchema.VariantLogicalType instance2 = HoodieSchema.VariantLogicalType.variant();
+
+    // Verify they are the exact same instance (reference equality)
+    assertSame(instance1, instance2);
+    assertTrue(instance1 == instance2);
+
+    // Verify same identity hash
+    assertEquals(System.identityHashCode(instance1), System.identityHashCode(instance2));
+
+    // Verify same object hash
+    assertEquals(instance1.hashCode(), instance2.hashCode());
+
+    // Verify the singleton is used in variant schemas
+    HoodieSchema.Variant variant1 = HoodieSchema.createVariant();
+    HoodieSchema.Variant variant2 = HoodieSchema.createVariant();
+
+    LogicalType logicalType1 = variant1.getAvroSchema().getLogicalType();
+    LogicalType logicalType2 = variant2.getAvroSchema().getLogicalType();
+
+    // Both should reference the same singleton instance
+    assertSame(logicalType1, logicalType2);
+    assertTrue(logicalType1 == logicalType2);
+    assertEquals(HoodieSchema.VariantLogicalType.variant(), logicalType1);
+
+    // Verify reference equality check works in fromAvroSchema
+    HoodieSchema parsedSchema = HoodieSchema.fromAvroSchema(variant1.getAvroSchema());
+    assertInstanceOf(HoodieSchema.Variant.class, parsedSchema);
+
+    // Verify the logical type name
+    assertEquals("variant", instance1.getName());
+  }
 }

hudi-spark-datasource/hudi-spark/src/test/scala/org/apache/spark/sql/hudi/feature/index/TestExpressionIndex.scala

@@ -159,7 +159,7 @@ class TestExpressionIndex extends HoodieSparkSqlTestBase with SparkAdapterSuppor
   }
 
   test("Test Create Expression Index Syntax") {
-    withTempDir { tmp =>
+    withTempDir { tmp =>2
       Seq("cow", "mor").foreach { tableType =>
         val databaseName = "default"
         val tableName = generateTableName