sksamuel / avro4s

Avro schema generation and serialization / deserialization



Build Status

Avro4s is a schema/class generation and serializing/deserializing library for Avro written in Scala. The objective is to allow seamless use with Scala without the need to to write boilerplate conversions yourself, and without the runtime overhead of reflection. Hence, this is a macro based library and generates code for use with Avro at compile time.

The features of the library are:

  • Schema generation from classes at compile time
  • Class generation from schemas at build time(through sbt)
  • Boilerplate free serialization of classes to Avro
  • Boilerplate free deserialization of Avro to classes


  • Fixed issue with serializing optional unions of Java enums
  • Added support for @AvroName annotation #159
  • Tuple2 and Tuple3 now supported in Schema generation #156
  • Added support for LocalDateTime
  • Added native support for Seq[Byte] mapped to Schema.Type.Bytes
  • Bumped Avro library to 1.8.2
  • @AvroFixed support for Strings
  • @AvroNamespace works on traits
  • Support overriding namespace per field #171
  • Fix NPE in support for Scala enumerations #158
  • Added scaling support for BigDecmials #142
  • Fixed bug with conversions to and from BigDecimal #127
  • Fixed support for value classes in scala 2.11
  • 1.6.4 - Added support for Vectors
  • 1.6.3 - Fixed issue with decimal schema using strings instead of ints for scale and precision
  • 1.6.2 - Removed unused imports, fixed compiler warnings
  • 1.6.1 - Fixed bug when using Option with a default value of None. Updated sbt generator. Added better way to discrimate between tpyes (Ilya Epifanov). Optimized reflection lookup @simonsouter.
  • 1.6.0 - Added support for Coproducts (see section on coproducts) @SohumB. Added support for value classes. Bumped 2.12 cross build to M4.
  • 1.5.6 - Added support for setting scale and precision on BigDecimal @blbradley
  • 1.5.5 - Skipped
  • 1.5.4 - Added support for recursive types @SohumB
  • 1.5.3 - Added support for Option[X] in default parameters @jecos
  • 1.5.2 - Added support for ints, longs, booleans, doubles, arrays, enums and maps in default parameters @whazenberg
  • 1.5.1 - Fixed macro bug introduced in 1.5.0 which broke macro generation.
  • 1.5.0 - Upgraded to Avro 1.8.1. Deprecated factory methods in io classes in favour of more explicit naming. Added caching for the macro generation (thanks @jtvoorde).
  • 1.4.3 - Added support for json serialization
  • 1.4.1 - Added support for parameter defaults
  • 1.4.0 - Added better support for enums. Added support for UUIDs. Rewrote the macro backend. Added better binary support.
  • 1.3.3 - Added missing support for deserializing byte arrays
  • 1.3.0 - Added support for Scala 2.12. Removed 2.10 cross build. Fixed issues with private vals. Added binary (no schema) output stream. Exposed RecordFormat[T] typeclass to enable easy conversion of T to/from an Avro Record.
  • 1.2.0 - Added support for properties, doc fields, and aliases. These are set via annotations.
  • 1.1.0 - Added JSON document to Avro schema converter
  • 1.0.0 - Migrated all macros to use Shapeless. Fixed some trickier nested case class issues. Simplified API. Added support for java enums.
  • 0.94.0 - Added support for writing/reading Either and Option in serializer/deserializer. Fixed bug with array serialization.
  • 0.93.0 - Added support for either and options in schema generator. Added support for aliases via scala annotation.
  • 0.92.0 - Added support for unions (and unions of nulls to Options) and enums to class generator.


Avro4s allows us to generate schemas directly from classes in a totally straightforward way. Let's define some classes.

case class Ingredient(name: String, sugar: Double, fat: Double)
case class Pizza(name: String, ingredients: Seq[Ingredient], vegetarian: Boolean, vegan: Boolean, calories: Int)

Next is to invoke the apply method of AvroSchema passing in the top level type. This will return an org.apache.avro.Schema instance, from which you can output, write to a file etc.

import com.sksamuel.avro4s.AvroSchema
val schema = AvroSchema[Pizza]

Which will output the following schema:

  "type" : "record",
  "name" : "Pizza",
  "namespace" : "com.sksamuel.avro4s.json",
  "fields" : [ {
    "name" : "name",
    "type" : "string"
  }, {
    "name" : "ingredients",
    "type" : {
      "type" : "array",
      "items" : {
        "type" : "record",
        "name" : "Ingredient",
        "fields" : [ {
          "name" : "name",
          "type" : "string"
        }, {
          "name" : "sugar",
          "type" : "double"
        }, {
          "name" : "fat",
          "type" : "double"
        } ]
  }, {
    "name" : "vegetarian",
    "type" : "boolean"
  }, {
    "name" : "vegan",
    "type" : "boolean"
  }, {
    "name" : "calories",
    "type" : "int"
  } ]

You can see that the schema generator handles nested case classes, sequences, primitives, etc. For a full list of supported object types, see the table later.

Recursive Schemas

Avro4s supports recursive schemas, but you will have to manually force the SchemaFor instance, instead of letting it be generated.

case class Recursive(payload: Int, next: Option[Recursive])

implicit val schemaFor = SchemaFor[Recursive]
val schema = AvroSchema[Recursive]

Input / Output


Avro4s allows us to easily serialize case classes using an instance of AvroOutputStream which we write to, and close, just like you would any regular output stream. An AvroOutputStream can be created from a File, Path, or by wrapping another OutputStream. When we create one, we specify the type of objects that we will be serializing. Eg, to serialize instances of our Pizza class:

import com.sksamuel.avro4s.AvroOutputStream

val pepperoni = Pizza("pepperoni", Seq(Ingredient("pepperoni", 12, 4.4), Ingredient("onions", 1, 0.4)), false, false, 98)
val hawaiian = Pizza("hawaiian", Seq(Ingredient("ham", 1.5, 5.6), Ingredient("pineapple", 5.2, 0.2)), false, false, 91)

val os =[Pizza](new File("pizzas.avro"))
os.write(Seq(pepperoni, hawaiian))


With avro4s we can easily deserialize a file back into Scala case classes. Given the pizzas.avro file we generated in the previous section on serialization, we will read this back in using the AvroInputStream class. We first create an instance of the input stream specifying the types we will read back, and the file. Then we call iterator which will return a lazy iterator (reads on demand) of the data in the file. In this example, we'll load all data at once from the iterator via toSet.

import com.sksamuel.avro4s.AvroInputStream

val is =[Pizza](new File("pizzas.avro"))
val pizzas = is.iterator.toSet

Will print out:

Pizza(pepperoni,List(Ingredient(pepperoni,12.2,4.4), Ingredient(onions,1.2,0.4)),false,false,500) Pizza(hawaiian,List(Ingredient(ham,1.5,5.6), Ingredient(pineapple,5.2,0.2)),false,false,500)

Binary Serializing

You can serialize to Binary using the AvroBinaryOutputStream which you can create directly or by using AvroOutputStream.binary. The difference with binary serialization is that the schema is not included in the output.


case class Composer(name: String, birthplace: String, compositions: Seq[String])
val ennio = Composer("ennio morricone", "rome", Seq("legend of 1900", "ecstasy of gold"))

val baos = new ByteArrayOutputStream()
val output = AvroOutputStream.binary[Composer](baos)
val bytes = baos.toByteArray

Binary Deserializing

You can deserialize Binary using the AvroBinaryInputStream which you can create directly or by using AvroInputStream.binary The difference with binary serialization is that the schema is not included in the data.

val in = new ByteArrayInputStream(bytes)
val input = AvroInputStream.binary[Composer](in)
val result = input.iterator.toSeq
result shouldBe Vector(ennio)

JSON Serializing

You can serialize to JSON using the AvroJsonOutputStream which you can create directly or by using AvroOutputStream.json


case class Composer(name: String, birthplace: String, compositions: Seq[String])
val ennio = Composer("ennio morricone", "rome", Seq("legend of 1900", "ecstasy of gold"))

val baos = new ByteArrayOutputStream()
val output = AvroOutputStream.json[Composer](baos)

JSON Deserializing

You can deserialize JSON using the AvroJsonInputStream which you can create directly or by using AvroInputStream.json

val json = "{\"name\":\"ennio morricone\",\"birthplace\":\"rome\",\"compositions\":[\"legend of 1900\",\"ecstasy of gold\"]}"
val in = new ByteInputStream(json.getBytes("UTF-8"), json.size)
val input = AvroInputStream.json[Composer](in)
val result = input.singleEntity
result shouldBe Success(ennio)

Conversions to/from GenericRecord

To interface with the Java API it is sometimes desirable to convert between your classes and the Avro GenericRecord type. You can do this easily in avro4s using the RecordFormat typeclass (this is what the input/output streams use behind the scenes). Eg,

To convert from a class into a record:

case class Composer(name: String, birthplace: String, compositions: Seq[String])
val ennio = Composer("ennio morricone", "rome", Seq("legend of 1900", "ecstasy of gold"))
val format = RecordFormat[Composer]
// record is of type GenericRecord
val record =

And to go from a record back into a type:

// given some record from earlier
val record = ...
val format = RecordFormat[Composer]
// is an instance of Composer
val ennio = format.from(record)

Set a schema's decimal scale, precision and rounding mode

Bring an implicit ScaleAndPrecisionAndRoundingMode into scope before using AvroSchema.

import com.sksamuel.avro4s.ScaleAndPrecisionAndRoundingMode

case class MyDecimal(d: BigDecimal)

implicit val sp = ScaleAndPrecisionAndRoundingMode(8, 20, RoundingMode.HALF_UP)
val schema = AvroSchema[MyDecimal]


Avro supports generalised unions, eithers of more than two values. To represent these in scala, we use shapeless.:+:, such that A :+: B :+: C :+: CNil represents cases where a type is A OR B OR C. See shapeless' documentation on coproducts for more on how to use coproducts.

Sealed hierarchies

Scala sealed traits/classes are supported both when it comes to schema generation and conversions to/from GenericRecord. Generally sealed hierarchies are encoded as unions - in the same way like Coproducts. Under the hood, shapeless Generic is used to derive Coproduct representation for sealed hierarchy.

When all descendants of sealed trait/class are singleton objects, optimized, enum-based encoding is used instead.

Type Mappings

import scala.collection.{Array, List, Seq, Iterable, Set, Map, Option, Either}
import shapeless.{:+:, CNil}
Scala Type Avro Type
Boolean boolean
Array[Byte] bytes
String string or fixed
Int int
Long long
BigDecimal decimal with default scale 2 and precision 8
Double double
Float float
java.util.UUID string
java.time.LocalDate string
java.time.LocalDateTime string
Java Enums enum
sealed trait T union
sealed trait with only case objects enum
Array[T] array
List[T] array
Seq[T] array
Iterable[T] array
Set[T] array
Map[String, T] map
Option[T] union:null,T
Either[L, R] union:L,R
A :+: B :+: C :+: CNil union:A,B,C
Option[Either[L, R]] union:null,L,R
Option[A :+: B :+: C :+: CNil] union:null,A,B,C
T record

Custom Type Mappings

It is very easy to add custom type mappings. To do this, you need to create instances of ToSchema, ToValue and FromValue typeclasses.

ToSchema is used to generate an Avro schema for a given JVM type. ToValue is used to convert an instance of a JVM type into an instance of the Avro type. And FromValue is used to convert an instance of the Avro type into the JVM type.

For example, to create a mapping for org.joda.time.DateTime that we wish to store as an ISO Date string, then we can do the following:

implicit object DateTimeToSchema extends ToSchema[DateTime] {
  override val schema: Schema = Schema.create(Schema.Type.STRING)

implicit object DateTimeToValue extends ToValue[DateTime] {
  override def apply(value: DateTime): String = ISODateTimeFormat.dateTime().print(value)

implicit object DateTimeFromValue extends FromValue[DateTime] {
  override def apply(value: Any, field: Field): DateTime = ISODateTimeFormat.dateTime().parseDateTime(value.toString())

These typeclasses must be implicit and in scope when you invoke AvroSchema or create an AvroInputStream/AvroOutputStream.

Selective Customisation

You can selectively customise the way Avro4s generates certain parts of your hierarchy, thanks to implicit precedence. Suppose you have the following classes:

case class Product(name: String, price: Price, litres: BigDecimal)
case class Price(currency: String, amount: BigDecimal)

And you want to selectively use different scale/precision for the price and litres quantities. You can do this by forcing the implicits in the corresponding companion objects.

object Price {
  implicit val sp = ScaleAndPrecision(10,2)
  implicit val schema = SchemaFor[Price]

object Product {
  implicit val sp = ScaleAndPrecision(8,4)
  implicit val schema = SchemaFor[Product]

This will result in a schema where both BigDecimal quantities have their own separate scale and precision.

Using avro4s in your project


compile 'com.sksamuel.avro4s:avro4s-core_2.11:xxx'


libraryDependencies += "com.sksamuel.avro4s" %% "avro4s-core" % "xxx"



Check the latest released version on Maven Central

Building and Testing

This project is built with SBT. So to build

sbt compile

And to test

sbt test


Contributions to avro4s are always welcome. Good ways to contribute include:

  • Raising bugs and feature requests
  • Fixing bugs and enhancing the DSL
  • Improving the performance of avro4s
  • Adding to the documentation