Query beans

Query beans are optional but they provide a nice way to write queries with type safe compile time checking. They are also easy to use and learn with IDE auto-completion on properties and expressions.

Query beans are generated using a Java annotation processing (APT) or Kotlin annotation processing (KAPT).

For each entity a query bean is generated with the same name but prefixed with Q. So for an entity bean called Customer there is a query bean generated called QCustomer.

When the entity bean model changes the query beans are regenerated and as developers we get compile time checking on our application queries - we get a compile time error if our queries are no longer valid for the model.

Contact contact =
  new QContact()                                // Contact query bean
    .email.equalTo("rob@foo.com")               // type safe expression
    .findOne();                                 // with IDE auto-completion
List<Customer> customers =
  new QCustomer()                                // Customer query bean
    .billingAddress.city.equalTo("Auckland")     // joins automatically added
    .contacts.isEmpty()                          // to support expressions

APT / KAPT (generating query beans)

To generate Java query beans we use the io.ebean:querybean-generator java annotation processor and to generate Kotlin query beans we use io.ebean:kotlin-querybean-generator.

Generating with Maven

Refer to docs / getting-started / maven for details of how to generate query beans using maven.

Generating with Gradle

Refer to docs / getting-started / gradle for details of how to generate query beans using gradle.


Note that prior to version 12.1.8 we need to edit a src/main/resources/ebean.mf manifest file to specify querybean-packages for the packages that should be enhanced for query beans. This is no longer needed from 12.1.8 onwards.

Example (Pre 12.1.8)
entity-packages: org.example.domain
transactional-packages: org.example
querybean-packages: org.example

Caveat - refactor rename

Refactor rename entity bean

When we refactor rename an entity bean in an IDE the query beans we should trigger a build all. This is because the query beans for the old bean names are kept around until the next build all.

Once we do a build all this will effectively delete the old query beans and generate the new ones. At this point we get compiler errors for the queries that we need to adjust (use the new query bean names).

Groovy Spock

When using query beans with Groovy perhaps in Spock tests we need to use them inside a @CompileStatic block.

static def findCount(String name) {
  return new QDbSample().name.eq(name).findCount()

Then we can use the helper method in our spock test like:

  // this line NPEs
  findCount('fred') == 1