Develop enterprise beans
One of two enterprise bean development scenarios is typically used with the product. The first is command-line using Ant, Make, Maven or similar tools. The second is an IDE-based development and build environment. The steps in this article focus on development without an IDE.
EJB 2.x beans only: Design a J2EE application and the enterprise beans that it needs.
- Before developing entity beans with container-managed persistence (CMP), read the topic Concurrency control.
EJB 3.x beans only: Design a Java EE application and the enterprise beans that it needs.
- Before developing entity beans with CMP, read the topic, Concurrency control. Keep in mind that EJB 3.x modules do not support entity beans. We must continue to place entity beans in your EJB 2.x-level modules.
The two basic approaches to select tools for developing enterprise beans are as follows:
- Using one of the available IDE tools that automatically generate significant parts of the
enterprise bean code and contain integrated tools for packaging and testing enterprise beans. The
Rational Application Developer product is the recommended
IDE.
Add install_root/dev/JavaEE/j2ee.jar to the IDE project build path to resolve compilation dependencies on the new EJB 3.x API classes. Code assist works when this JAR file is added to the project build path. If we define a server (see J2EE Perspective), point the server to the product installation directory. When we create a Java EE related project in Rational Application Developer, the project automatically adds install_root/dev/JavaEE/j2ee.jar. to the project build path.
- If we have decided to develop enterprise beans without an IDE, we need at least an ASCII text
editor. We can also use a Java development tool that does not
support enterprise bean development. We can then use tools available in the Java SDK (SDK) and in this product to assemble, test, and
deploy the beans.
Like the assembly tool, a standard Java EE command-line build environment requires some change to use the EJB 3.x modules. As with previous Java EE application development patterns, we must include the j2ee.jar file located in the install_root/dev/JavaEE directory on the compiler class path. An example of a command-line build environment using Ant is located in the install_root/samples/src/TechSamp directory.
The following steps primarily support the second approach, development without an IDE.
Tasks
- If necessary, migrate any pre-existing code to the required version of the EJB
specification.
Applications written to the EJB specification versions 1.1, 2.0, and 2.1 can run unchanged in the EJB 3.x container. See the topic Migrating enterprise bean code to the supported specification.
- Write and compile the components of the enterprise bean.
- At a minimum, a session bean developed with the EJB 3.x specifications requires a bean class.
- At a minimum, an EJB 1.1 session bean requires a bean class, a home interface, and a remote
interface. An EJB 1.1 entity bean requires a bean class, a primary-key class, a home interface, and
a remote interface.
- At a minimum, an EJB 2.x session bean requires a bean class, a home or local home interface, and
a remote or local interface. An EJB 2.x entity bean requires a bean class, a primary-key class, a
remote home or local home interface, and a remote or local interface. The types of interfaces go
together: If we implement a local interface, we must also define a local home interface.
The primary-key class can be unknown. See the topic Unknown primary-key class for more information.
- A message-driven bean requires only a bean class.
- At a minimum, a session bean developed with the EJB 3.x specifications requires a bean class.
- For each entity bean, complete work to handle persistence operations.
For EJB 3.x modules, consider using the Java Persistence API (JPA) specification to develop plain old Java Object (POJO) persistent entities. Review the topic Java Persistence API for more information. If we choose to develop entity beans to earlier EJB specifications, follow these steps:
- Create a database schema for the entity bean persistent data.
- For entity beans with CMP, we must store the bean persistent data in one of the supported databases. The assembly tool automatically generates SQL code for creating database tables for CMP entity beans. If our CMP beans require complex database mappings, IBM recommends that we use Rational Application Developer to generate code for the database tables. For more information about using the assembly tools, see the assembly tool information center.
- For entity beans with bean-managed persistence (BMP), we can create the database and database table using the database tools or use an existing database and database table.
For more information about creating databases and database tables, review the database documentation.
- (CMP entity beans for EJB 2.x only)
Define finder queries with EJB Query Language (EJB QL).
With EJB QL, we define finders in terms of CMP fields and container-managed relationships, as follows:
- Public finders are visible in the bean home interface. Implemented in the bean class, they return only remote interfaces and collection types.
- Private finders, expressed as SELECT statements, are used only within the bean class. They can return both local and remote interfaces, dependent values, other CMP field types, and collection types.
- (CMP entity beans for EJB 1.1 only: an IBM
extension) Create a finder helper interface for each CMP entity bean containing specialized
finder methods (other than the findByPrimaryKey method).
Logic other than the findByPrimaryKey method is required for each finder method contained in the home interface of an entity bean with CMP:
- The logic must be defined in a public interface named NameBeanFinderHelper, where Name is the name of the enterprise bean, for example, AccountBeanFinderHelper.
- The logic must be contained in a String constant named findMethodName WhereClause, where findMethodName is the name of the finder method. The String constant can contain zero or more question marks (?) that are replaced from left to right with the value of the finder method arguments when that method is called.
- Create a database schema for the entity bean persistent data.
Example: Using a read-only entity bean
This usage scenario and example shows how to write an EJB application that uses a read-only entity bean.
- Usage scenario
-
A customer has a database of catalog pricing and shipping rate information that is updated daily no later than 10:00 PM local time (22:00 in 24-hour format). They want to write an EJB application that has read-only access to this data. That is, this application never updates the pricing database. Updating is done through some other application.
- Example
-
The customer's entity bean local interface might be:
public interface ItemCatalogData extends EJBLocalObject { public int getItemPrice(); public int getShippingCost(int destinationCode); }The code in the stateless SessionBean method (assume it is a TxRequired) that invokes this EntityBean to figure out the total cost including shipping, would look like:
..... // Some transactional steps occur prior to this point, such as removing the item from // inventory, etc. // Now obtain the price of this item and start to calculate the total cost to the purchaser ItemCatalogData theItemData = (ItemCatalogData) ItemCatalogDataHome.findByPrimaryKey(theCatalogNumber); int totalcost = theItemData.getItemPrice(); // ... some other processing, etc. in the interim // ... // ... // Add the shipping costs totalcost = totalcost + theItemData.getShippingCost(theDestinationPostalCode);
At application assembly time, the customer sets the EJB caching parameters for this bean as follows:
- ActivateAt = ONCE
- LoadAt = DAILY
- ReloadInterval = 2200
Deprecated feature: The reloadInterval and reloadingEnabled
attributes of the IBM deployment descriptor extensions,
including both the WAR file extension (WEB-INF/ibm-web-ext.xmi) and the application extension
(META-INF/ibm-application-ext.xmi) were deprecated.depfeat
On the first call to the getItemPrice() method after 22:00 each night, the EJB container reloads the pricing information from the database. If the clock strikes 22:00 between the call to getItemPrice() and getShippingCost(), the getShippingCost() method still returns the value it had before any changes to the database that might have occurred at 22:00, since the first method invocation in this transaction occurred before 22:00. Thus, the item price and shipping cost used remain in sync with each other.
What to do next
Assemble the beans in one or more EJB modules. See the topic Assembling EJB modules, or Assembling EJB 3.x modules if we are using EJB 3.x beans.
Subtopics
- Develop message-driven beans
- Enterprise bean development best practices
- WebSphere extensions to the Enterprise JavaBeans specification
- Setting the run time for batched commands with JVM arguments
- Setting the run time for deferred create with JVM arguments
- Setting persistence manager cache invalidation
- Setting the system property to enable remote EJB clients to receive nested or root-cause exceptions
- Unknown primary-key class
- Develop message-driven beans
- Enterprise bean development best practices
- WebSphere extensions to the Enterprise JavaBeans specification
- Setting the run time for batched commands with JVM arguments
- Setting the run time for deferred create with JVM arguments
- Setting persistence manager cache invalidation
- Setting the system property to enable remote EJB clients to receive nested or root-cause exceptions
- Unknown primary-key class
Related: