Enumeratum is a type-safe and powerful enumeration implementation for Scala that offers exhaustive pattern match warnings,
integrations with popular Scala libraries, and idiomatic usage that won't break your IDE. It aims to be similar enough
to Scala's built in Enumeration
to be easy-to-use and understand while offering more flexibility, type-safety (see this blog
post describing erasure on Scala's Enumeration
), and
richer enum values without having to maintain your own collection of values.
Enumeratum has the following niceties:
- Zero dependencies
- Performant: Faster than
Enumeration
in the standard library (see benchmarks) - Allows your Enum members to be full-fledged normal objects with methods, values, inheritance, etc.
ValueEnum
s that map to various primitive values and have compile-time uniqueness constraints.- Idiomatic: you're very clearly still writing Scala, and no funny colours in your IDE means less cognitive overhead for your team
- Simplicity; most of the complexity in this lib is in its macro, and the macro is fairly simple conceptually
- No usage of reflection at runtime. This may also help with performance but it means Enumeratum is compatible with ScalaJS and other environments where reflection is a best effort (such as Android)
- No usage of
synchronized
, which may help with performance and deadlocks prevention - All magic happens at compile-time so you know right away when things go awry
- Comprehensive automated testing to make sure everything is in tip-top shape
Enumeratum is published for Scala 2.10.x, 2.11.x, and 2.12.x as well as ScalaJS.
Integrations are available for:
- Play: JVM only
- Play JSON: JVM (included in Play integration but also available separately) and ScalaJS
- Circe: JVM and ScalaJS
- UPickle: JVM and ScalaJS
- ReactiveMongo BSON: JVM only
- Argonaut: JVM and ScalaJS
- Json4s: JVM only
- ScalaCheck: JVM and ScalaJS
- Slick: JVM only
- Quill: JVM and ScalaJS
- Quick start
- SBT
- Usage
- More examples
- Enum 1. Manual override of name 2. Mixins to override the name
- ValueEnum
- ScalaJS
- Play integration
- Play JSON integration
- Circe integration
- UPickle integration
- ReactiveMongo BSON integration
- Argonaut integration
- Json4s integration
- Slick integration
- ScalaCheck
- Quill integration
- Benchmarking
- Publishing
In build.sbt
, set the Enumeratum version in a variable (for the latest version, set val enumeratumVersion =
the version you see
in the Maven badge above).
libraryDependencies ++= Seq(
"com.beachape" %% "enumeratum" % enumeratumVersion
)
Enumeratum has different integrations that can be added to your build à la carte. For more info, see the respective sections in the Table of Contents
Using Enumeratum is simple. Just declare your own sealed trait or class A
that extends EnumEntry
and implement it as case objects inside
an object that extends from Enum[A]
as shown below.
import enumeratum._
sealed trait Greeting extends EnumEntry
object Greeting extends Enum[Greeting] {
/*
`findValues` is a protected method that invokes a macro to find all `Greeting` object declarations inside an `Enum`
You use it to implement the `val values` member
*/
val values = findValues
case object Hello extends Greeting
case object GoodBye extends Greeting
case object Hi extends Greeting
case object Bye extends Greeting
}
// Object Greeting has a `withName(name: String)` method
Greeting.withName("Hello")
// => res0: Greeting = Hello
Greeting.withName("Haro")
// => java.lang.IllegalArgumentException: Haro is not a member of Enum (Hello, GoodBye, Hi, Bye)
// A safer alternative would be to use `withNameOption(name: String)` method which returns an Option[Greeting]
Greeting.withNameOption("Hello")
// => res1: Option[Greeting] = Some(Hello)
Greeting.withNameOption("Haro")
// => res2: Option[Greeting] = None
// It is also possible to use strings case insensitively
Greeting.withNameInsensitive("HeLLo")
// => res3: Greeting = Hello
Greeting.withNameInsensitiveOption("HeLLo")
// => res4: Option[Greeting] = Some(Hello)
// Uppercase-only strings may also be used
Greeting.withNameUppercaseOnly("HELLO")
// => res5: Greeting = Hello
Greeting.withNameUppercaseOnlyOption("HeLLo")
// => res6: Option[Greeting] = None
// Similarly, lowercase-only strings may also be used
Greeting.withNameLowercaseOnly("hello")
// => res7: Greeting = Hello
Greeting.withNameLowercaseOnlyOption("hello")
// => res8: Option[Greeting] = Some(Hello)
Note that by default, findValues
will return a Seq
with the enum members listed in written-order (relevant if you want to
use the indexOf
method).
For an interactive demo, checkout this Scastie snippet.
Continuing from the enum declared in the quick-start section:
import Greeting._
def tryMatching(v: Greeting): Unit = v match {
case Hello => println("Hello")
case GoodBye => println("GoodBye")
case Hi => println("Hi")
}
/**
Pattern match warning ...
<console>:24: warning: match may not be exhaustive.
It would fail on the following input: Bye
def tryMatching(v: Greeting): Unit = v match {
*/
Greeting.indexOf(Bye)
// => res2: Int = 3
The name is taken from the toString
method of the particular
EnumEntry
. This behavior can be changed in two ways.
The first way to change the name behaviour is to manually override the def entryName: String
method.
import enumeratum._
sealed abstract class State(override val entryName: String) extends EnumEntry
object State extends Enum[State] {
val values = findValues
case object Alabama extends State("AL")
case object Alaska extends State("AK")
// and so on and so forth.
}
import State._
State.withName("AL")
The second way to override the name behaviour is to mixin the stackable traits provided for common string
conversions, Snakecase
, UpperSnakecase
, CapitalSnakecase
, Hyphencase
, UpperHyphencase
, CapitalHyphencase
, Dotcase
, UpperDotcase
, CapitalDotcase
, Words
, UpperWords
, CapitalWords
, Camelcase
, LowerCamelcase
, Uppercase
, Lowercase
, and Uncapitalised
.
import enumeratum._
import enumeratum.EnumEntry._
sealed trait Greeting extends EnumEntry with Snakecase
object Greeting extends Enum[Greeting] {
val values = findValues
case object Hello extends Greeting
case object GoodBye extends Greeting
case object ShoutGoodBye extends Greeting with Uppercase
}
Greeting.withName("hello")
Greeting.withName("good_bye")
Greeting.withName("SHOUT_GOOD_BYE")
Asides from enumerations that resolve members from String
names, Enumeratum also supports ValueEnum
s, enums that resolve
members from simple values like Int
, Long
, Short
, Char
, Byte
, and String
(without support for runtime transformations).
These enums are not modelled after Enumeration
from standard lib, and therefore have the added ability to make sure, at compile-time,
that multiple members do not share the same value.
import enumeratum.values._
sealed abstract class LibraryItem(val value: Int, val name: String) extends IntEnumEntry
case object LibraryItem extends IntEnum[LibraryItem] {
case object Book extends LibraryItem(value = 1, name = "book")
case object Movie extends LibraryItem(name = "movie", value = 2)
case object Magazine extends LibraryItem(3, "magazine")
case object CD extends LibraryItem(4, name = "cd")
// case object Newspaper extends LibraryItem(4, name = "newspaper") <-- will fail to compile because the value 4 is shared
/*
val five = 5
case object Article extends LibraryItem(five, name = "article") <-- will fail to compile because the value is not a literal
*/
val values = findValues
}
assert(LibraryItem.withValue(1) == LibraryItem.Book)
LibraryItem.withValue(10) // => java.util.NoSuchElementException:
Restrictions
ValueEnum
s must have their value members implemented as literal values.
In a ScalaJS project, add the following to build.sbt
:
libraryDependencies ++= Seq(
"com.beachape" %%% "enumeratum" % enumeratumVersion
)
As expected, usage is exactly the same as normal Scala.
The enumeratum-play
project is published separately and gives you access to various tools
to help you avoid boilerplate in your Play project.
For enumeratum with full Play support:
libraryDependencies ++= Seq(
"com.beachape" %% "enumeratum-play" % enumeratumPlayVersion
)
Note that as of version 1.4.0, enumeratum-play
for Scala 2.11 is compatible with Play 2.5+ while 2.10 is compatible with
Play 2.4.x. Versions prior to 1.4.0 are compatible with 2.4.x.
The included PlayEnum
trait is probably going to be the most interesting as it includes a bunch
of built-in implicits like Json formats, Path bindables, Query string bindables, and Form field support.
For example:
package enums._
import enumeratum._
sealed trait Greeting extends EnumEntry
object Greeting extends PlayEnum[Greeting] {
val values = findValues
case object Hello extends Greeting
case object GoodBye extends Greeting
case object Hi extends Greeting
case object Bye extends Greeting
}
/*
Then make sure to import your PlayEnums into your routes in your Build.scala
or build.sbt so that you can use them in your routes file.
`routesImport += "enums._"`
*/
// You can also use the String Interpolating Routing DSL:
import play.api.routing.sird._
import play.api.routing._
import play.api.mvc._
Router.from {
case GET(p"/hello/${Greeting.fromPath(greeting)}") => Action {
Results.Ok(s"$greeting")
}
}
There are IntPlayEnum
, LongPlayEnum
, and ShortPlayEnum
traits for use with IntEnumEntry
, LongEnumEntry
, and
ShortEnumEntry
respectively that provide Play-specific implicits as with normal PlayEnum
. For example:
import enumeratum.values._
sealed abstract class PlayLibraryItem(val value: Int, val name: String) extends IntEnumEntry
case object PlayLibraryItem extends IntPlayEnum[PlayLibraryItem] {
// A good mix of named, unnamed, named + unordered args
case object Book extends PlayLibraryItem(value = 1, name = "book")
case object Movie extends PlayLibraryItem(name = "movie", value = 2)
case object Magazine extends PlayLibraryItem(3, "magazine")
case object CD extends PlayLibraryItem(4, name = "cd")
val values = findValues
}
import play.api.libs.json.{ JsNumber, JsString, Json => PlayJson }
PlayLibraryItem.values.foreach { item =>
assert(PlayJson.toJson(item) == JsNumber(item.value))
}
PlayEnum
extends the trait PlayFormFieldEnum
wich offers formField
for mapping within a play.api.data.Form
object.
import play.api.data.Form
import play.api.data.Forms._
object GreetingForm {
val form = Form(
mapping(
"name" -> nonEmptyText,
"greeting" -> Greeting.formField
)(Data.apply)(Data.unapply)
)
case class Data(
name: String,
greeting: Greeting)
}
Another alternative (if for example your Enum
can't extend PlayEnum
or PlayFormFieldEnum
) is to create an implicit Format
and bring it into scope using Play's of
, i.e.
import play.api.data.Form
import play.api.data.Forms._
object Formats {
implicit val greetingFormat = enumeratum.Forms.format(Greeting)
}
object GreetingForm {
import Formats._
val form = Form(
mapping(
"name" -> nonEmptyText,
"greeting" -> of[Greeting]
)(Data.apply)(Data.unapply)
)
case class Data(
name: String,
greeting: Greeting)
}
The enumeratum-play-json
project is published separately and gives you access to Play's auto-generated boilerplate
for JSON serialization in your Enum's.
libraryDependencies ++= Seq(
"com.beachape" %% "enumeratum-play-json" % enumeratumPlayJsonVersion
)
Note that as of version 1.4.0, enumeratum-play
for Scala 2.11 is compatible with Play 2.5+ while 2.10 is compatible with
Play 2.4.x. Versions prior to 1.4.0 are compatible with 2.4.x.
There are also PlayInsensitiveJsonEnum
, PlayLowercaseJsonEnum
, and PlayUppercaseJsonEnum
traits for use. For example:
import enumeratum.{ PlayJsonEnum, Enum, EnumEntry }
sealed trait Greeting extends EnumEntry
object Greeting extends Enum[Greeting] with PlayJsonEnum[Greeting] {
val values = findValues
case object Hello extends Greeting
case object GoodBye extends Greeting
case object Hi extends Greeting
case object Bye extends Greeting
}
There are IntPlayJsonEnum
, LongPlayJsonEnum
, and ShortPlayJsonEnum
traits for use with IntEnumEntry
, LongEnumEntry
, and
ShortEnumEntry
respectively. For example:
import enumeratum.values._
sealed abstract class JsonDrinks(val value: Short, name: String) extends ShortEnumEntry
case object JsonDrinks extends ShortEnum[JsonDrinks] with ShortPlayJsonValueEnum[JsonDrinks] {
case object OrangeJuice extends JsonDrinks(value = 1, name = "oj")
case object AppleJuice extends JsonDrinks(value = 2, name = "aj")
case object Cola extends JsonDrinks(value = 3, name = "cola")
case object Beer extends JsonDrinks(value = 4, name = "beer")
val values = findValues
}
import play.api.libs.json.{ JsNumber, JsString, Json => PlayJson, JsSuccess }
// Use to deserialise numbers to enum members directly
JsonDrinks.values.foreach { drink =>
assert(PlayJson.toJson(drink) == JsNumber(drink.value))
}
assert(PlayJson.fromJson[JsonDrinks](JsNumber(3)) == JsSuccess(JsonDrinks.Cola))
assert(PlayJson.fromJson[JsonDrinks](JsNumber(19)).isError)
To use enumeratum with Circe:
libraryDependencies ++= Seq(
"com.beachape" %% "enumeratum-circe" % enumeratumCirceVersion
)
To use with ScalaJS:
libraryDependencies ++= Seq(
"com.beachape" %%% "enumeratum-circe" % enumeratumCirceVersion
)
import enumeratum._
sealed trait ShirtSize extends EnumEntry
case object ShirtSize extends Enum[ShirtSize] with CirceEnum[ShirtSize] {
case object Small extends ShirtSize
case object Medium extends ShirtSize
case object Large extends ShirtSize
val values = findValues
}
import io.circe.Json
import io.circe.syntax._
ShirtSize.values.foreach { size =>
assert(size.asJson == Json.fromString(size.entryName))
}
import enumeratum.values._
sealed abstract class CirceLibraryItem(val value: Int, val name: String) extends IntEnumEntry
case object CirceLibraryItem extends IntEnum[CirceLibraryItem] with IntCirceEnum[CirceLibraryItem] {
// A good mix of named, unnamed, named + unordered args
case object Book extends CirceLibraryItem(value = 1, name = "book")
case object Movie extends CirceLibraryItem(name = "movie", value = 2)
case object Magazine extends CirceLibraryItem(3, "magazine")
case object CD extends CirceLibraryItem(4, name = "cd")
val values = findValues
}
import io.circe.Json
import io.circe.syntax._
CirceLibraryItem.values.foreach { item =>
assert(item.asJson == Json.fromInt(item.value))
}
To use enumeratum with uPickle:
libraryDependencies ++= Seq(
"com.beachape" %% "enumeratum-upickle" % enumeratumUPickleVersion
)
To use with ScalaJS:
libraryDependencies ++= Seq(
"com.beachape" %%% "enumeratum-upickle" % enumeratumUPickleVersion
)
UPickleEnum
works pretty much the same as CirceEnum
and PlayJsonEnum
variants, so we'll skip straight to the
ValueEnum
integration.
import enumeratum.values._
sealed abstract class ContentType(val value: Long, name: String) extends LongEnumEntry
case object ContentType
extends LongEnum[ContentType]
with LongUPickleEnum[ContentType] {
val values = findValues
case object Text extends ContentType(value = 1L, name = "text")
case object Image extends ContentType(value = 2L, name = "image")
case object Video extends ContentType(value = 3L, name = "video")
case object Audio extends ContentType(value = 4L, name = "audio")
}
import upickle.default.{ readJs, writeJs, Reader, Writer }
enum.values.foreach { entry =>
val written = writeJs(entry)
assert(readJs(written) == entry)
}
The enumeratum-reactivemongo-bson
project is published separately and gives you access to ReactiveMongo's auto-generated boilerplate
for BSON serialization in your Enum's.
libraryDependencies ++= Seq(
"com.beachape" %% "enumeratum-reactivemongo-bson" % enumeratumReactiveMongoVersion
)
For example:
import enumeratum.{ ReactiveMongoBsonEnum, Enum, EnumEntry }
sealed trait Greeting extends EnumEntry
object Greeting extends Enum[Greeting] with ReactiveMongoBsonEnum[Greeting] {
val values = findValues
case object Hello extends Greeting
case object GoodBye extends Greeting
case object Hi extends Greeting
case object Bye extends Greeting
}
There are IntReactiveMongoBsonValueEnum
, LongReactiveMongoBsonValueEnum
, and ShortReactiveMongoBsonValueEnum
traits for use with IntEnumEntry
, LongEnumEntry
, and
ShortEnumEntry
respectively. For example:
import enumeratum.values._
sealed abstract class BsonDrinks(val value: Short, name: String) extends ShortEnumEntry
case object BsonDrinks extends ShortEnum[BsonDrinks] with ShortReactiveMongoBsonValueEnum[BsonDrinks] {
case object OrangeJuice extends BsonDrinks(value = 1, name = "oj")
case object AppleJuice extends BsonDrinks(value = 2, name = "aj")
case object Cola extends BsonDrinks(value = 3, name = "cola")
case object Beer extends BsonDrinks(value = 4, name = "beer")
val values = findValues
}
import reactivemongo.bson._
// Use to deserialise numbers to enum members directly
BsonDrinks.values.foreach { drink =>
val writer = implicitly[BSONWriter[BsonDrinks, BSONValue]]
assert(writer.write(drink) == BSONInteger(drink.value))
}
val reader = implicitly[BSONReader[BSONValue, BsonDrinks]]
assert(reader.read(BSONInteger(3)) == BsonDrinks.Cola)
To use enumeratum with Argonaut:
libraryDependencies ++= Seq(
"com.beachape" %% "enumeratum-argonaut" % enumeratumArgonautVersion
)
import enumeratum._
sealed trait TrafficLight extends EnumEntry
object TrafficLight extends Enum[TrafficLight] with ArgonautEnum[TrafficLight] {
case object Red extends TrafficLight
case object Yellow extends TrafficLight
case object Green extends TrafficLight
val values = findValues
}
import argonaut._
import Argonaut._
TrafficLight.values.foreach { entry =>
assert(entry.asJson == entry.entryName.asJson)
}
import enumeratum.values._
sealed abstract class ArgonautDevice(val value: Short) extends ShortEnumEntry
case object ArgonautDevice
extends ShortEnum[ArgonautDevice]
with ShortArgonautEnum[ArgonautDevice] {
case object Phone extends ArgonautDevice(1)
case object Laptop extends ArgonautDevice(2)
case object Desktop extends ArgonautDevice(3)
case object Tablet extends ArgonautDevice(4)
val values = findValues
}
import argonaut._
import Argonaut._
ArgonautDevice.values.foreach { item =>
assert(item.asJson == item.value.asJson)
}
To use enumeratum with Json4s:
libraryDependencies ++= Seq(
"com.beachape" %% "enumeratum-json4s" % enumeratumJson4sVersion
)
import enumeratum._
sealed trait TrafficLight extends EnumEntry
object TrafficLight extends Enum[TrafficLight] /* nothing extra here */ {
case object Red extends TrafficLight
case object Yellow extends TrafficLight
case object Green extends TrafficLight
val values = findValues
}
import org.json4s.DefaultFormats
implicit val formats = DefaultFormats + Json4s.serializer(TrafficLight)
import enumeratum.values._
sealed abstract class Device(val value: Short) extends ShortEnumEntry
case object Device
extends ShortEnum[Device] /* nothing extra here */ {
case object Phone extends Device(1)
case object Laptop extends Device(2)
case object Desktop extends Device(3)
case object Tablet extends Device(4)
val values = findValues
}
import org.json4s.DefaultFormats
implicit val formats = DefaultFormats + Json4s.serializer(Device)
To use enumeratum with ScalaCheck:
libraryDependencies ++= Seq(
"com.beachape" %% "enumeratum-scalacheck" % enumeratumScalacheckVersion
)
Given the enum declared in the quick-start section, you can get an Arbitrary[Greeting]
(to generate instances of Greeting
) and a Cogen[Greeting]
(to generate instances of Greeting => (A: Arbitrary)
) by importing generators in the scope of your tests:
import enumeratum.scalacheck._
Similarly, you can get Arbitrary
and Cogen
instances for every ValueEnum
subtype by importing generators in the scope of your tests:
import enumeratum.values.scalacheck._
To use enumeratum with Quill:
libraryDependencies ++= Seq(
"com.beachape" %% "enumeratum-quill" % enumeratumQuillVersion
)
To use with ScalaJS:
libraryDependencies ++= Seq(
"com.beachape" %%% "enumeratum-quill" % enumeratumQuillVersion
)
import enumeratum._
sealed trait ShirtSize extends EnumEntry
case object ShirtSize extends Enum[ShirtSize] with QuillEnum[ShirtSize] {
case object Small extends ShirtSize
case object Medium extends ShirtSize
case object Large extends ShirtSize
val values = findValues
}
case class Shirt(size: ShirtSize)
import io.getquill._
lazy val ctx = new PostgresJdbcContext(SnakeCase, "ctx")
import ctx._
ctx.run(query[Shirt].insert(_.size -> lift(ShirtSize.Small: ShirtSize)))
ctx.run(query[Shirt]).foreach(println)
- Note that a type ascription to the
EnumEntry
trait (eg.ShirtSize.Small: ShirtSize
) is required when binding hardcodedEnumEntry
s
import enumeratum._
sealed abstract class ShirtSize(val value: Int) extends IntEnumEntry
case object ShirtSize extends IntEnum[ShirtSize] with IntQuillEnum[ShirtSize] {
case object Small extends ShirtSize(1)
case object Medium extends ShirtSize(2)
case object Large extends ShirtSize(3)
val values = findValues
}
case class Shirt(size: ShirtSize)
import io.getquill._
lazy val ctx = new PostgresJdbcContext(SnakeCase, "ctx")
import ctx._
ctx.run(query[Shirt].insert(_.size -> lift(ShirtSize.Small: ShirtSize)))
ctx.run(query[Shirt]).foreach(println)
- Note that a type ascription to the
ValueEnumEntry
abstract class (eg.ShirtSize.Small: ShirtSize
) is required when binding hardcodedValueEnumEntry
s quill-cassandra
currently does not supportShortEnum
andByteEnum
(see getquill/quill#1009)quill-orientdb
currently does not supportByteEnum
(see getquill/quill#1029)
In order to use your enumeratum Enums in Slick tables as columns, you will
need to construct instances of MappedColumnType
and make them available
where you define and query your slick tables. In order to more easily
construct these instances, the enumeratum-slick integration provides a trait
enumeratum.SlickEnumSupport
. This trait provides a method mappedColumnTypeForEnum
(and variants) for constructing a mapped column type for your enum. For example
if you want to use Enum[Greeting]
in your slick table, mix in SlickEnumSupport
where you define your table.
trait GreetingRepository extends SlickEnumSupport {
val profile: slick.jdbc.Profile
implicit lazy val greetingMapper = mappedColumnTypeForEnum(Greeting)
class GreetingTable(tag: Tag) extends Table[(String, Greeting)](tag, "greeting") {
def id = column[String]("id", O.PrimaryKey)
def greeting = column[Greeting]("greeting") // Maps to a varchar/text column
def * = (id, greeting)
}
If you want to represent a ValueEnum
by its value
rather than its string
name, simply mix in SlickValueEnumSupport
and proceed mostly as above:
implicit lazy val libraryItemMapper = mappedColumnTypeForIntEnum(LibraryItem)
...
def item = column[LibraryItem]("LIBRARY_ITEM") // Maps to a numeric column
An alternate approach which is useful when mappers need to be shared across repositories (perhaps for something common like a "Status" enum) is to define your mappers in a module on their own, then make use of them in your repositories:
trait CommonMappers extends SlickEnumSupport {
val profile: Profile
implicit lazy val statusMapper = mappedColumnTypeForEnum(Status)
...
}
trait UserRepository extends CommonMappers {
val profile: Profile
class UserTable(tag: Tag) extends Table[UserRow](tag, "user") {
...
def status = column[Status]("status")
...
}
}
Note that because your enum values are singleton objects, you may get errors when you try to use them in Slick queries like the following:
.filter(_.trafficLight === TrafficLight.Red)
This is because TrafficLight.Red
in the above example is inferred to
be of its unique type (TrafficLight.Red
) rather than TrafficLight
,
thus causing a failure to find your mapping. In order to fix this,
simply assist the compiler by ascribing the type to be TrafficLight
:
.filter(_.trafficLight === (TrafficLight.Red: TrafficLight))
A way around this if you find the type expansion offensive is to define val accessors for your enum entries that are typed as the parent type. You can do this inside your Enums companion object or more locally:
val red: TrafficLight = Red // Not red: TrafficLight.Red = Red
val yellow: TrafficLight = Yellow
val green: TrafficLight = Green
...
.filter(_.trafficLight === red)
If you want to use slick interpolated SQL queries you can use the provided
constructors to instantiate instances of GetResult[_]
and SetParameter[_]
for your enum:
import SlickEnumPlainSqlSupport._
Or mix it in...
trait Foo extends SlickEnumPlainSqlSupport {
...
}
Then define your instances:
implicit val greetingGetResult = getResultForEnum(Greeting)
implicit val greetingOptionGetResult = optionalGetResultForEnum(Greeting)
implicit val greetingSetParameter = setParameterForEnum(Greeting)
implicit val greetingOptionSetParameter = optionalSetParameterForEnum(Greeting)
Benchmarking is in the unpublished benchmarking
project. It uses JMH and you can run them in the sbt console by issuing the following command from your command line:
sbt +benchmarking/'jmh:run -i 10 -wi 10 -f3 -t 1'
The above command will run JMH benchmarks against different versions of Scala. Leave off +
to run against the main/latest supported version of Scala.
On my late 2013 MBP using Java8 on OSX El Capitan:
[info] Benchmark Mode Cnt Score Error Units
[info] EnumBenchmarks.indexOf avgt 30 11.628 ± 0.190 ns/op
[info] EnumBenchmarks.withNameDoesNotExist avgt 30 1809.194 ± 33.113 ns/op
[info] EnumBenchmarks.withNameExists avgt 30 13.540 ± 0.374 ns/op
[info] EnumBenchmarks.withNameOptionDoesNotExist avgt 30 5.999 ± 0.037 ns/op
[info] EnumBenchmarks.withNameOptionExists avgt 30 9.662 ± 0.232 ns/op
[info] StdLibEnumBenchmarks.withNameDoesNotExist avgt 30 1921.690 ± 78.898 ns/op
[info] StdLibEnumBenchmarks.withNameExists avgt 30 56.517 ± 1.161 ns/op
[info] values.ValueEnumBenchmarks.withValueDoesNotExist avgt 30 1950.291 ± 29.292 ns/op
[info] values.ValueEnumBenchmarks.withValueExists avgt 30 4.009 ± 0.062 ns/op
[info] values.ValueEnumBenchmarks.withValueOptDoesNotExist avgt 30 5.285 ± 0.063 ns/op
[info] values.ValueEnumBenchmarks.withValueOptExists avgt 30 6.621 ± 0.084 ns/op
Other than the methods that throw NoSuchElementException
s, performance is in the 10ns range (taking into account JMH overhead of roughly 2-3ns), which
is acceptable for almost all use-cases. PRs that promise to increase performance are expected to be demonstrably faster.
Also, Enumeratum's withName
is faster than the standard library's Enumeration
, by around 4x in the case where an entry exists with the given name.
My guess is this is because Enumeratum doesn't use any synchronized
calls or volatile
annotations. It is also faster in the case where there is no
corresponding name, but not by a significant amount, perhaps because the high cost of throwing an exception masks any benefits.
Projects are published independently of each other.
JVM + ScalaJS projects should have an aggregate project to make it easy to publish them, e.g. for enumeratum-circe
:
$ sbt "project circe-aggregate" +clean +publish-signed
Should publish all needed artefacts. Note that sbt circe-aggregate/publish-signed
will not work (ScalaJS gets skipped).