§Using the Ebean ORM

§Configuring Ebean

Play comes with the Ebean ORM. To enable it, add javaEbean to your
dependencies :

val appDependencies = Seq(
  javaEbean
)

then add the following line to conf/application.conf:

ebean.default="models.*"

This defines a default Ebean server, using the default data source, which must be properly configured. You can actually create as many Ebean servers you need, and explicitly define the mapped class for each server.

ebean.orders="models.Order,models.OrderItem"
ebean.customers="models.Customer,models.Address"

In this example, we have access to two Ebean servers - each using its own database.

Each ebean. config line (as above) can map any classes that Ebean may be interested in registering (eg. @Entity/Model classes, @Embeddables, custom ScalarTypes and CompoundTypes, BeanPersistControllers, BeanPersistListeners, BeanFinders, ServerConfigStartups, etc). These can be individually listed separated by commas, and/or you can use the wildcard .*. For example, models.* registers with Ebean all classes within the models package that Ebean can make use of.

To customise the underlying Ebean Server configuration, you can either add a conf/ebean.properties file, or create an instance of the ServerConfigStartup interface to programmatically manipulate the Ebean ServerConfig before the server is initialised.

As an example, the fairly common problem of reducing the Sequence Batch Size in order to minimise sequence gaps, could be solved quite simply with a class like this:

package models;

import com.avaje.ebean.config.ServerConfig;
import com.avaje.ebean.event.ServerConfigStartup;

public class MyServerConfigStartup implements ServerConfigStartup {
    @Override
    public void onStart(ServerConfig serverConfig) {
        serverConfig.setDatabaseSequenceBatchSize(1);
    }
}

Note that Ebean will also make use of a conf/orm.xml file (if present), to configure <entity-mappings>.

For more information about Ebean, see the Ebean documentation.

§Using the play.db.ebean.Model superclass

Play defines a convenient superclass for your Ebean model classes. Here is a typical Ebean class, mapped in Play:

package models;

import java.util.*;
import javax.persistence.*;

import play.db.ebean.*;
import play.data.format.*;
import play.data.validation.*;

@Entity 
public class Task extends Model {

  @Id
  @Constraints.Min(10)
  public Long id;
  
  @Constraints.Required
  public String name;
  
  public boolean done;
  
  @Formats.DateTime(pattern="dd/MM/yyyy")
  public Date dueDate = new Date();
  
  public static Finder<Long,Task> find = new Finder<Long,Task>(
    Long.class, Task.class
  ); 

}

Play has been designed to generate getter/setter automatically, to ensure compatibility with libraries that expect them to be available at runtime (ORM, Databinder, JSON Binder, etc). If Play detects any user-written getter/setter in the Model, it will not generate getter/setter in order to avoid any conflict.

Caveats:

(1) Because Ebean class enhancement occurs after compilation, do not expect Ebean-generated getter/setters to be available at compilation time. If you’d prefer to code with them directly, either add the getter/setters explicitly yourself, or ensure that your model classes are compiled before the remainder of your project, eg. by putting them in a separate subproject.

(2) Enhancement of direct Ebean field access (enabling lazy loading) is only applied to Java classes, not to Scala. Thus, direct field access from Scala source files (including standard Play templates) does not invoke lazy loading, often resulting in empty (unpopulated) entity fields. To ensure the fields get populated, either (a) manually create getter/setters and call them instead, or (b) ensure the entity is fully populated before accessing the fields.

As you can see, we’ve added a find static field, defining a Finder for an entity of type Task with a Long identifier. This helper field is then used to simplify querying our model:

// Find all tasks
List<Task> tasks = Task.find.all();
    
// Find a task by ID
Task anyTask = Task.find.byId(34L);

// Delete a task by ID
Task.find.ref(34L).delete();

// More complex task query
List<Task> tasks = find.where()
    .ilike("name", "%coco%")
    .orderBy("dueDate asc")
    .findPagingList(25)
    .setFetchAhead(false)
    .getPage(1);

§Transactional actions

By default Ebean will use transactions. However this transactions will be created before and commited or rollbacked after every single query, update, create or delete, as you can see here:

//Created implicit transaction
List<User> users =   
            Ebean.find(User.class)  
                .join("customer")  
                .where().eq("state", UserState.ACTIVE)  
                .findList();  
//Transaction commited or rollbacked

//Created implicit transaction
Ebean.save(user);
//Transaction commited or rollbacked

So, if you want to do more than one action in the same transaction you can use TxRunnable and TxCallable:

// run in Transactional scope...  
Ebean.execute(new TxRunnable() {  
  public void run() {  
      
    // code running in "REQUIRED" transactional scope  
    // ... as "REQUIRED" is the default TxType  
    System.out.println(Ebean.currentTransaction());  
      
    // find stuff...  
    User user = Ebean.find(User.class, 1);  
    ...  
      
    // save and delete stuff...  
    Ebean.save(user);  
    Ebean.delete(order);  
    ...  
  }  
});

You can also, if the class is a POJO one, annotate your action method with @play.db.ebean.Transactional to compose your action method with an Action that will automatically manage a transaction:

@Transactional
public static Result save() {
  ...
}

Or if you want a more traditional approach you can begin, commit and rollback transactions explicitly:

Ebean.beginTransaction();  
try {  
    // fetch some stuff...  
    User u = Ebean.find(User.class, 1);  
    ...  
  
    // save or delete stuff...  
    Ebean.save(u);  
    ...  
  
    Ebean.commitTransaction();  
      
} finally {  
    Ebean.endTransaction();  
}  

Next: Integrating with JPA