Get a typed value from an Avro GenericRecord

Question

Given a GenericRecord, what is the recommended way to retrieve a typed value, as opposed to an Object? Are we expected to cast the values, and if so what is the mapping from Avro types to Java types? For example, Avro Array == Java Collection; and Avro String == Java Utf8.

Since every GenericRecord contains its schema, I was hoping for a type-safe way to retrieve values.

Answer 1

Avro has eight primitive types and five complex types (excluding unions which are a combination of other types). The following table maps these 13 Avro types to their input interfaces (the Java types which can be put into a GenericRecord) and their output implementations (the concrete Java types which are returned by a get from a GenericRecord). The values apply to Avro 1.7.7.

╔═══════════╦════════════════════════╦═══════════════════════════╗ ║ Avro Type ║ Input Interface ║ Output Implementation ║ ╠═══════════╬════════════════════════╬═══════════════════════════╣ ║ null ║ ║ null ║ ║ boolean ║ java.lang.Boolean ║ java.lang.Boolean ║ ║ int ║ java.lang.Integer ║ java.lang.Integer ║ ║ long ║ java.lang.Long ║ java.lang.Long ║ ║ float ║ java.lang.Float ║ java.lang.Float ║ ║ double ║ java.lang.Double ║ java.lang.Double ║ ║ bytes ║ java.nio.ByteBuffer ║ java.nio.HeapByteBuffer ║ ║ string ║ java.lang.CharSequence ║ org.apache.avro.util.Utf8 ║ ║ record ║ *.GenericRecord ║ *.GenericData$Record ║ ║ enum ║ java.lang.CharSequence ║ *.GenericData$EnumSymbol ║ ║ array ║ java.util.Collection ║ *.GenericData$Array ║ ║ map ║ java.util.Map ║ java.util.HashMap ║ ║ fixed ║ *.GenericFixed ║ *.GenericData$Fixed ║ ╚═══════════╩════════════════════════╩═══════════════════════════╝ * == org.apache.avro.generic

---

In Avro 1.8.0, the enum type requires a GenericEnumSymbol. It no longer accepts CharSequence.