Documentation

You are viewing the documentation for the 2.6.0-M4 development release. The latest stable release series is 3.0.x.

§The Play cache API

Caching data is a typical optimization in modern applications, and so Play provides a global cache.

An important point about the cache is that it behaves just like a cache should: the data you just stored may just go missing.

For any data stored in the cache, a regeneration strategy needs to be put in place in case the data goes missing. This philosophy is one of the fundamentals behind Play, and is different from Java EE, where the session is expected to retain values throughout its lifetime.

The default implementation of the Cache API uses EHCache.

§Importing the Cache API

Add cache into your dependencies list. For example, in build.sbt:

libraryDependencies ++= Seq(
  cache,
  ...
)

§Accessing the Cache API

The cache API is defined by the AsyncCacheApi and SyncCacheApi traits, depending on whether you want an asynchronous or synchronous implementation, and can be injected into your component like any other dependency. For example:

import play.api.cache._
import play.api.mvc._
import javax.inject.Inject

class Application @Inject() (cache: AsyncCacheApi) extends Controller {

}

Note: The API is intentionally minimal to allow several implementation to be plugged in. If you need a more specific API, use the one provided by your Cache plugin.

Using this simple API you can store data in cache:

val result: Future[Done] = cache.set("item.key", connectedUser)

And then retrieve it later:

val futureMaybeUser: Future[Option[User]] = cache.get[User]("item.key")

There is also a convenient helper to retrieve from cache or set the value in cache if it was missing:

val futureUser: Future[User] = cache.getOrElseUpdate[User]("item.key") {
  User.findById(connectedUser)
}

You can specify an expiry duration by passing a duration, by default the duration is infinite:

import scala.concurrent.duration._

val result: Future[Done] = cache.set("item.key", connectedUser, 5.minutes)

To remove an item from the cache use the remove method:

val removeResult: Future[Done] = cache.remove("item.key")

Note that the SyncCacheApi has the same API, except it returns the values directly instead of using futures.

§Accessing different caches

It is possible to access different caches. The default cache is called play, and can be configured by creating a file called ehcache.xml. Additional caches may be configured with different configurations, or even implementations.

If you want to access multiple different ehcache caches, then you’ll need to tell Play to bind them in application.conf, like so:

play.cache.bindCaches = ["db-cache", "user-cache", "session-cache"]

By default, Play will try to create these caches for you. If you would like to define them yourself in ehcache.xml, you can set:

play.cache.createBoundCaches = false

Now to access these different caches, when you inject them, use the NamedCache qualifier on your dependency, for example:

import play.api.cache._
import play.api.mvc._
import javax.inject.Inject

class Application @Inject()(
    @NamedCache("session-cache") sessionCache: AsyncCacheApi
) extends Controller {

}

§Caching HTTP responses

You can easily create smart cached actions using standard Action composition.

Note: Play HTTP Result instances are safe to cache and reuse later.

The Cached class helps you build cached actions.

import play.api.cache.Cached
import javax.inject.Inject

class Application @Inject() (cached: Cached) extends Controller {

}

You can cache the result of an action using a fixed key like "homePage".

def index = cached("homePage") {
  Action {
    Ok("Hello world")
  }
}

If results vary, you can cache each result using a different key. In this example, each user has a different cached result.

def userProfile = WithAuthentication(_.session.get("username")) { userId =>
  cached(req => "profile." + userId) {
    Action.async {
      User.find(userId).map { user =>
        Ok(views.html.profile(user))
      }
    }
  }
}

§Control caching

You can easily control what you want to cache or what you want to exclude from the cache.

You may want to only cache 200 Ok results.

def get(index: Int) = cached.status(_ => "/resource/"+ index, 200) {
  Action {
    if (index > 0) {
      Ok(Json.obj("id" -> index))
    } else {
      NotFound
    }
  }
}

Or cache 404 Not Found only for a couple of minutes

def get(index: Int) = {
  val caching = cached
    .status(_ => "/resource/"+ index, 200)
    .includeStatus(404, 600)

  caching {
    Action {
      if (index % 2 == 1) {
        Ok(Json.obj("id" -> index))
      } else {
        NotFound
      }
    }
  }
}

§Custom implementations

It is possible to provide a custom implementation of the cache API that either replaces, or sits along side the default implementation.

To replace the default implementation, you’ll need to disable the default implementation by setting the following in application.conf:

play.modules.disabled += "play.api.cache.EhCacheModule"

You can then implement AsyncCacheApi and bind it in the DI container. You can also bind SyncCacheApi to DefaultSyncCacheApi, which simply wraps the async implementation.

To provide an implementation of the cache API in addition to the default implementation, you can either create a custom qualifier, or reuse the NamedCache qualifier to bind the implementation.

Next: Calling WebServices