I would like to know the difference between creating an DAO Object with constructor like :
myDaoObject = new MyDaoObject();
and creating it by EJB injection :
#EJB
MyDaoObject myDaoObject;
Is there a difference in the running and/or performance ?
thanks.
Well...you really can't create EJBs with constructor because you would lose the functionality offered by the container (dependency injection, pooling, calling of #PostConstruct, transactions, ...). So only correct way is
#EJB
MyDaoObject myDaoObject;
P.S. Or, in case you are using CDI, #Inject instead of #EJB
Related
I am new to spring and not able to understand when to instantiate the class with new operator and when by using spring container.
example i found a code
import org.springframework.context.ApplicationContext;
import org.springframework.context.support.ClassPathXmlApplicationContext;
public class Test {
public static void main(String[] args) {
ApplicationContext ctx=new ClassPathXmlApplicationContext("applicationContext.xml");
EmployeeDao dao=(EmployeeDao)ctx.getBean("edao");
int status=dao.saveEmployee(new Employee(102,"Amit",35000));
System.out.println(status);
int status=dao.updateEmployee(new Employee(102,"Sonoo",15000));
System.out.println(status);
Employee e=new Employee();
e.setId(102);
int status=dao.deleteEmployee(e);
System.out.println(status);
}
}
i am using jdbc template. i have a doubt why we have new to instatiate employee class instead we should have used (Employee)ctx.getBean("employee"). using new operator would create dependency?
Pls help
From ProSpring book,
"Of the applications that we have built using Spring,
the only objects that are consistently not managed by Spring are domain objects. (Even though in Spring
it’s possible to have Spring manage domain objects by applying the #Component annotation to the classes
and assigning them with prototype scope, most of the time we will choose to manage domain objects
within the application.) The reason for this is that, practically, Spring does not need to be involved with
domain objects. Generally, you create many instances of your domain objects using the new() operator
and perform processing either in the service or data access layer. Although Spring also supports the
injection of new instance of domain objects every time it was requested (by using the bean scope
prototype), generally developers will not adopt this approach since typically domain objects do not take
advantage of Dependency Injection, because they generally have few dependencies outside of the DOM
itself, and they don’t require much configuration."
So, yes you could create beans of type Employee but then they wil have to be defined as prototype as be default all beans in spring are singletons. because there can be many employees and their objects, you obviously cannot have Employee as a singleton. but your concentration as to the utilization of spring features should be much more than managing domain objects.
The Spring IOC container is at the core of the Spring Framework. The container will create the objects, wire them together, configure them, and manage their complete lifecycle from creation till destruction.The IoC container gets informations/metadata either from XML or by Java annotations, or by Java code.
Yeah you can create Employee bean in your spring context xml file and make sure scope is prototype as
<bean id="employee" class="com.mycompany.Employee" scope="prototype"/>
so that on every request of Employee bean, you will get a new object Employee.
There are other scopes of beans
singleton : This scopes the bean definition to a single instance per Spring IoC container (default).
request : On every HTTP request a new bean will created and delivered.
prototype : on every request(api request) a new bean will be created.
session : This scope a bean definition to an HTTP session.
Note : It is not recomemded to create beans for database model class.because your persistent layer will take care of it.
I am using EJB in order to take advantages of:
Concurrent (instead of creating 2 threads, I divided the work into 2
EJB beans).
Pooling (I use stateless EJB a lot and I love the idea that the pool
contains a specific number of bean). This way, I am not afraid of
running out of memory. Memory usage is more predictable).
Asynchronous processing (all I need is just an annotation).
Well, the problem is I am using it with MongoDB so I don't need any transaction. I can use #TransactionAttribute(TransactionAttributeType.NOT_SUPPORTED) and #TransactionAttribute(TransactionAttributeType.NEVER) annotations but... it means I must specify it everywhere?
Is there anyway to disable EJB transaction by default?
In an EJB 3.0 container, annotate your EJB (or EJB method) with:
#Stateless
#TransactionManagement(TransactionManagementType.BEAN)
#TransactionAttribute(value=TransactionAttributeType.NEVER)
public class YourBean
for BEAN management. For CONTAINER management instead:
#Stateless
#TransactionManagement(TransactionManagementType.CONTAINER)
#TransactionAttribute(value=TransactionAttributeType.NEVER)
public class YourBean
The default value is managed by the container but if you dont specify nothing to do i think you solve your problem.
Or annotate all the Ejb to don´t support transaction
#Stateless
#TransactionManagement(TransactionManagementType.NEVER)
public class YourBean
Remember that the ejb transactions are executed in a hierarchical way, ie if the first method being invoked does not support methods "children methods" are handled in the same way
I have Enterprise Java Bean which is Statefull and it holds current user instance
I want to get this instance from few different baking beans (SessionScoped managed beans) but when Im using:
#EJB
UserSessionBean usb;
(...)
usb.getUser();
I am getting null pointer exceptions in the ManagedBean (seems that every managedbean is getting new instance of UserSessionBean EJB. Why is that? I thought one instance of that bean would be shared among all Beans for that session...
if UserSessionBean is the implementation class or interface?
I think you should use the interface to create an instance of your EJB.
also take a look at #statefull and #singlton annotations.
Is it enough to convert a POJO like Util class to an EJB session bean by putting an annotation (#Stateless or #Stateful) and using injected EntityManager in it?
Yes, #Stateless in enough. Your bean will then become an EJB bean.
The only other requirement is that you can't create such a bean with new. You have to inject it using #EJB in another managed bean (JSF managed bean, Servlet, etc). Or if you aren't yet in any kind of managed bean, you can bootstrap a bean using a JNDI lookup.
Also, EJBs indeed greatly reduce the boilerplate code of starting and committing transactions when working with JPA.
Well It is enough but still few things need to be taken Care,
1) Mark your Entity manager and other new variables to Transient if POJO is used to persist some object.
2) It is better not to do so, as If u need to make it as EJB better to Create Some New Class for it as it is suggested way to not to create complexity.
I want to call to a method from an EJB in the same instant in which one deploys itself, without using a servlet.
Thanks.
David.
There seems to be no life-cycle methods defined by the EJB spec for this purpose. Individual vendors may provide extensions to allow this. For example Startup Beans in WebSphere would be a place to put the invocation logic you want.
Using techniques such as a static method seem slightly dangerous in that we don't know whether all dependency injection is complete before that static method fires, and hence whether you can safely use the business methods of the EJB.
Persoanlly, if I needed to be portable I would bite the bullet and use a servlet. It costs very little.
Try doing your initialization within a static block. This will run once when the classloader loads the class.
static { System.out.println("static"); }
The PostConstruct hook is right for that.
Find more info on about PostConstruct here:
in the javadoc
lifecycle of EJBs in the JavaEE 5 tutorial
Let's finish with a quick example:
#Stateless
public class TestEJB implements MyEJBInterface{
#PostConstruct
public void doThatAfterInitialization() {
//put your code here to be executed after creation and initialization of your bean
}
}
Static initializer blocks are technically not illegal in EJB but they are used to execute code before any constructor (which might be a problem) when instantiating a class. They are typically used to initialize static fields which may be illegal in EJB if they are not read only. So, what about using ejbCreate(), setSessionContext() or setEntityContext() methods instead (not even sure this would be appropriate without more details on the problem you are trying to solve)?
The EJB container, for a #Singleton bean, shall create the instance of the bean as soon as the application is deploy if it is annotated #Startup.
That will, of course, fire up static initialization blocks, the constructor, dependency injection setters, #PostConstruct methods etc.
Here is the appropriate referente to the Java EE 6 Tutorial.