Added java string option to ReflectionSchemaBuilder

sksamuel · Apr 14, 2024 · aff0edd · aff0edd
1 parent 9f61760
commit aff0edd
Show file tree

Hide file tree

Showing 4 changed files with 32 additions and 3 deletions.
diff --git a/centurion-avro/src/main/kotlin/com/sksamuel/centurion/avro/encoders/strings.kt b/centurion-avro/src/main/kotlin/com/sksamuel/centurion/avro/encoders/strings.kt
@@ -11,7 +11,7 @@ object StringEncoder : Encoder<String> {
       return when (schema.type) {
          Schema.Type.BYTES -> ByteStringEncoder.encode(schema, value)
          Schema.Type.FIXED -> FixedStringEncoder.encode(schema, value)
-         Schema.Type.STRING -> when (schema.getObjectProp(GenericData.STRING_PROP)) {
+         Schema.Type.STRING -> when (schema.getProp(GenericData.STRING_PROP)) {
             "String" -> value
             else -> UTF8StringEncoder.encode(schema, value)
          }

diff --git a/...on-avro/src/main/kotlin/com/sksamuel/centurion/avro/generation/ReflectionSchemaBuilder.kt b/...on-avro/src/main/kotlin/com/sksamuel/centurion/avro/generation/ReflectionSchemaBuilder.kt
@@ -2,11 +2,14 @@ package com.sksamuel.centurion.avro.generation
 
 import org.apache.avro.Schema
 import org.apache.avro.SchemaBuilder
+import org.apache.avro.generic.GenericData
 import kotlin.reflect.KClass
 import kotlin.reflect.KType
 import kotlin.reflect.full.memberProperties
 
-class ReflectionSchemaBuilder {
+class ReflectionSchemaBuilder(
+   private val useJavaString: Boolean = false,
+) {
 
    fun schema(kclass: KClass<*>): Schema {
 
@@ -27,6 +30,8 @@ class ReflectionSchemaBuilder {
 
       return when (val classifier = type.classifier) {
          String::class -> typeBuilder.stringType()
+            .also { if (useJavaString) GenericData.setStringType(it, GenericData.StringType.String) }
+
          Boolean::class -> typeBuilder.booleanType()
          Int::class -> typeBuilder.intType()
          Long::class -> typeBuilder.longType()

diff --git a/centurion-avro/src/test/kotlin/com/sksamuel/centurion/avro/encoders/Benchmarks.kt b/centurion-avro/src/test/kotlin/com/sksamuel/centurion/avro/encoders/Benchmarks.kt
@@ -22,7 +22,9 @@ enum class UserType { User, Admin }
 fun main() {
 
    val user = User(Random.nextLong(), "sammy mcsamface", "[email protected]", Random.nextLong(), UserType.Admin)
-   val schema = ReflectionSchemaBuilder().schema(User::class)
+   val schema = ReflectionSchemaBuilder(true).schema(User::class)
+
+   GenericData.setStringType(schema, GenericData.StringType.String)
 
    repeat(5) {
       val time = measureTime {

diff --git a/centurion-avro/src/test/kotlin/com/sksamuel/centurion/avro/encoders/StringEncoderTest.kt b/centurion-avro/src/test/kotlin/com/sksamuel/centurion/avro/encoders/StringEncoderTest.kt
@@ -0,0 +1,22 @@
+package com.sksamuel.centurion.avro.encoders
+
+import io.kotest.core.spec.style.FunSpec
+import io.kotest.matchers.types.shouldBeTypeOf
+import org.apache.avro.SchemaBuilder
+import org.apache.avro.generic.GenericData
+import org.apache.avro.util.Utf8
+
+class StringEncoderTest : FunSpec({
+
+   test("should use java string when prop is set") {
+      val schema = SchemaBuilder.builder().stringType()
+      GenericData.setStringType(schema, GenericData.StringType.String)
+      StringEncoder.encode(schema, "hello").shouldBeTypeOf<String>()
+   }
+
+   test("should use Utf8 when prop is not set") {
+      val schema = SchemaBuilder.builder().stringType()
+      StringEncoder.encode(schema, "hello").shouldBeTypeOf<Utf8>()
+   }
+
+})