I want to publish an event if and only if there were changes to the DB. I'm running under #Transaction is Spring context and I come up with this check:
Session session = entityManager.unwrap(Session.class);
session.isDirty();
That seems to fail for new (Transient) objects:
#Transactional
public Entity save(Entity newEntity) {
Entity entity = entityRepository.save(newEntity);
Session session = entityManager.unwrap(Session.class);
session.isDirty(); // <-- returns `false` ):
return entity;
}
Based on the answer here https://stackoverflow.com/a/5268617/672689 I would expect it to work and return true.
What am I missing?
UPDATE
Considering #fladdimir answer, although this function is called in a transaction context, I did add the #Transactional (from org.springframework.transaction.annotation) on the function. but I still encounter the same behaviour. The isDirty is returning false.
Moreover, as expected, the new entity doesn't shows on the DB while the program is hold on breakpoint at the line of the session.isDirty().
UPDATE_2
I also tried to change the session flush modes before calling the repo save, also without any effect:
session.setFlushMode(FlushModeType.COMMIT);
session.setHibernateFlushMode(FlushMode.MANUAL);
First of all, Session.isDirty() has a different meaning than what I understood. It tells if the current session is holding in memory queries which still haven't been sent to the DB. While I thought it tells if the transaction have changing queries. When saving a new entity, even in transaction, the insert query must be sent to the DB in order to get the new entity id, therefore the isDirty() will always be false after it.
So I ended up creating a class to extend SessionImpl and hold the change status for the session, updating it on persist and merge calls (the functions hibernate is using)
So this is the class I wrote:
import org.hibernate.HibernateException;
import org.hibernate.internal.SessionCreationOptions;
import org.hibernate.internal.SessionFactoryImpl;
import org.hibernate.internal.SessionImpl;
public class CustomSession extends SessionImpl {
private boolean changed;
public CustomSession(SessionFactoryImpl factory, SessionCreationOptions options) {
super(factory, options);
changed = false;
}
#Override
public void persist(Object object) throws HibernateException {
super.persist(object);
changed = true;
}
#Override
public void flush() throws HibernateException {
changed = changed || isDirty();
super.flush();
}
public boolean isChanged() {
return changed || isDirty();
}
}
In order to use it I had to:
extend SessionFactoryImpl.SessionBuilderImpl to override the openSession function and return my CustomSession
extend SessionFactoryImpl to override the withOptions function to return the extended SessionFactoryImpl.SessionBuilderImpl
extend AbstractDelegatingSessionFactoryBuilderImplementor to override the build function to return the extended SessionFactoryImpl
implement SessionFactoryBuilderFactory to implement getSessionFactoryBuilder to return the extended AbstractDelegatingSessionFactoryBuilderImplementor
add org.hibernate.boot.spi.SessionFactoryBuilderFactory file under META-INF/services with value of my SessionFactoryBuilderFactory implementation full class name (for the spring to be aware of it).
UPDATE
There was a bug with capturing the "merge" calls (as tremendous7 comment), so I end up capturing the isDirty state before any flush, and also checking it once more when checking isChanged()
The following is a different way you might be able to leverage to track dirtiness.
Though architecturally different than your sample code, it may be more to the point of your actual goal (I want to publish an event if and only if there were changes to the DB).
Maybe you could use an Interceptor listener to let the entity manager do the heavy lifting and just TELL you what's dirty. Then you only have to react to it, instead of prod it to sort out what's dirty in the first place.
Take a look at this article: https://www.baeldung.com/hibernate-entity-lifecycle
It has a lot of test cases that basically check for dirtiness of objects being saved in various contexts and then it relies on a piece of code called the DirtyDataInspector that effectively listens to any items that are flagged dirty on flush and then just remembers them (i.e. keeps them in a list) so the unit test cases can assert that the things that SHOULD have been dirty were actually flushed as dirty.
The dirty data inspector code is on their github. Here's the direct link for ease of access.
Here is the code where the interceptor is applied to the factory so it can be effective. You might need to write this up in your injection framework accordingly.
The code for the Interceptor it is based on has a TON of lifecycle methods you can probably exploit to get the perfect behavior for "do this if there was actually a dirty save that occured".
You can see the full docs of it here.
We do not know your complete setup, but as #Christian Beikov suggested in the comment, is it possible that the insertion was already flushed before you call isDirty()?
This would happen when you called repository.save(newEntity) without a running transaction, since the SimpleJpaRepository's save method is annotated itself with #Transactional:
#Transactional
#Override
public <S extends T> S save(S entity) {
...
}
This will wrap the call in a new transaction if none is already active, and flush the insertion to the DB at the end of the transaction just before the method returns.
You might choose to annotate the method where you call save and isDirty with #Transactional, so that the transaction is created when your method is called, and propagated to the repository call. This way the transaction would not be committed when the save returns, and the session would still be dirty.
(edit, just for completeness: in case of using an identity ID generation strategy, the insertion of newly created entity is flushed during a repository's save call to generate the ID, before the running transaction is committed)
Related
I have this class and I tought three ways to handle detached entity state in case of persistence exceptions (which are handled elsewhere):
#ManagedBean
#ViewScoped
public class EntityBean implements Serializable
{
#EJB
private PersistenceService service;
private Document entity;
public void update()
{
// HANDLING 1. ignore errors
service.transact(em ->
{
entity = em.merge(entity);
// some other code that modifies [entity] properties:
// entity.setCode(...);
// entity.setResposible(...);
// entity.setSecurityLevel(...);
}); // an exception may be thrown on method return (rollback),
// but [entity] has already been reassigned with a "dirty" one.
//------------------------------------------------------------------
// HANDLING 2. ensure entity is untouched before flush is ok
service.transact(em ->
{
Document managed = em.merge(entity);
// some other code that modifies [managed] properties:
// managed.setCode(...);
// managed.setResposible(...);
// managed.setSecurityLevel(...);
em.flush(); // an exception may be thrown here (rollback)
// forcing method exit without [entity] being reassigned.
entity = managed;
}); // an exception may be thrown on method return (rollback),
// but [entity] has already been reassigned with a "dirty" one.
//------------------------------------------------------------------
// HANDLING 3. ensure entity is untouched before whole transaction is ok
AtomicReference<Document> reference = new AtomicReference<>();
service.transact(em ->
{
Document managed = em.merge(entity);
// some other code that modifies [managed] properties:
// managed.setCode(...);
// managed.setResposible(...);
// managed.setSecurityLevel(...);
reference.set(managed);
}); // an exception may be thrown on method return (rollback),
// and [entity] is safe, it's not been reassigned yet.
entity = reference.get();
}
...
}
PersistenceService#transact(Consumer<EntityManager> consumer) can throw unchecked exceptions.
The goal is to maintain the state of the entity aligned with the state of the database, even in case of exceptions (prevent entity to become "dirty" after transaction fail).
Method 1. is obviously naive and doesn't guarantee coherence.
Method 2. asserts that nothing can go wrong after flushing.
Method 3. prevents the new entity assigment if there's an exception in the whole transaction
Questions:
Is method 3. really safer than method 2.?
Are there cases where an exception is thrown between flush [excluded] and commit [included]?
Is there a standard way to handle this common problem?
Thank you
Note that I'm already able to rollback the transaction and close the EntityManager (PersistenceService#transact will do it gracefully), but I need to solve database state and the business objects do get out of sync. Usually this is not a problem. In my case this is the problem, because exceptions are usually generated by BeanValidator (those on JPA side, not on JSF side, for computed values that depends on user inputs) and I want the user to input correct values and try again, without losing the values he entered before.
Side note: I'm using Hibernate 5.2.1
this is the PersistenceService (CMT)
#Stateless
#Local
public class PersistenceService implements Serializable
{
#PersistenceContext
private EntityManager em;
#TransactionAttribute(TransactionAttributeType.REQUIRED)
public void transact(Consumer<EntityManager> consumer)
{
consumer.accept(em);
}
}
#DraganBozanovic
That's it! Great explanation for point 1. and 2.
I'd just love you to elaborate a little more on point 3. and give me some advice on real-world use case.
However, I would definitely not use AtomicReference or similar cumbersome constructs. Java EE, Spring and other frameworks and application containers support declaring transactional methods via annotations: Simply use the result returned from a transactional method.
When you have to modify a single entity, the transactional method would just take the detached entity as parameter and return the updated entity, easy.
public Document updateDocument(Document doc)
{
Document managed = em.merge(doc);
// managed.setXxx(...);
// managed.setYyy(...);
return managed;
}
But when you need to modify more than one in a single transaction, the method can become a real pain:
public LinkTicketResult linkTicket(Node node, Ticket ticket)
{
LinkTicketResult result = new LinkTicketResult();
Node managedNode = em.merge(node);
result.setNode(managedNode);
// modify managedNode
Ticket managedTicket = em.merge(ticket);
result.setTicket(managedTicket);
// modify managedTicket
Remark managedRemark = createRemark(...);
result.setRemark(managedemark);
return result;
}
In this case, my pain:
I have to create a dedicated transactional method (maybe a dedicated #EJB too)
That method will be called only once (will have just one caller) - is a "one-shot" non-reusable public method. Ugly.
I have to create the dummy class LinkTicketResult
That class will be instantiated only once, in that method - is "one-shot"
The method could have many parameters (or another dummy class LinkTicketParameters)
JSF controller actions, in most cases, will just call a EJB method, extract updated entities from returned container and reassign them to local fields
My code will be steadily polluted with "one-shotters", too many for my taste.
Probably I'm not seeing something big that's just in front of me, I'll be very grateful if you can point me in the right direction.
Is method 3. really safer than method 2.?
Yes. Not only is it safer (see point 2), but it is conceptually more correct, as you change transaction-dependent state only when you proved that the related transaction has succeeded.
Are there cases where an exception is thrown between flush [excluded] and commit [included]?
Yes. For example:
LockMode.OPTIMISTIC:
Optimistically assume that transaction will not experience contention
for entities. The entity version will be verified near the transaction
end.
It would be neither performant nor practically useful to check optimistick lock violation during each flush operation within a single transaction.
Deferred integrity constraints (enforced at commit time in db). Not used often, but are an illustrative example for this case.
Later maintenance and refactoring. You or somebody else may later introduce additional changes after the last explicit call to flush.
Is there a standard way to handle this common problem?
Yes, I would say that your third approach is the standard one: Use the results of a complete and successful transaction.
However, I would definitely not use AtomicReference or similar cumbersome constructs. Java EE, Spring and other frameworks and application containers support declaring transactional methods via annotations: Simply use the result returned from a transactional method.
Not sure if this is entirely to the point, but there is only one way to recover after exceptions: rollback and close the EM. From https://docs.jboss.org/hibernate/entitymanager/3.6/reference/en/html/transactions.html#transactions-basics-issues
An exception thrown by the Entity Manager means you have to rollback
your database transaction and close the EntityManager immediately
(discussed later in more detail). If your EntityManager is bound to
the application, you have to stop the application. Rolling back the
database transaction doesn't put your business objects back into the
state they were at the start of the transaction. This means the
database state and the business objects do get out of sync. Usually
this is not a problem, because exceptions are not recoverable and you
have to start over your unit of work after rollback anyway.
-- EDIT--
Also see http://piotrnowicki.com/2013/03/jpa-and-cmt-why-catching-persistence-exception-is-not-enough/
ps: downvote is not mine.
I have a method that is defined as #transactional. In fact I have a method calling a method that calls a method and all three are #transactional. The transactional logic worked fine, until I pulled a few methods out into an abstract class for some code reuse, which appears to have broken my logic somehow.
The transactional method is from an abstract class, here is a partial snippet of the relevant parts (I have to rewrite this by hand so forgive me for typos):
public abstract class ReadWriteService<ReadEntityTempalte extends IEntity, WriteEntityTemplate extends IEntity>
//extends jpaRepository, created using #enableJpaRepositories
private searchRepository<WriteEntityTemplate, String> writeRepository;
#PersistenceContext
private EntityManager em;
#transactional
public ReadEntityTemplate save(final WriteEntityTemplate entity){
if(entity == null) return null;
WriteEntityTemplate returnValue = writeRepository_.save(entity);
postSave(returnValue); //checks our security logic
flush();
ReadEntityTemplate returnEntity = find(returnValue.getId());
//required to detect changes made to the view by our save
em.refresh(returnEntity);
}
It's written this way because we are using views so the return value may be modified in the find() to the view. This logic worked in the past, and still works for a number of calls.
The method that fails is:
#Override
#transational
public void configure(EntityFileConfig config) throws ClassNotFoundException{
//load config from file
for(EntityConfig entityConfig: entityConfigs){
EntityType entityType=EntityTypeService_.find(entityConfig.getKey());
if(entityType==null){
entityType = EntityType.createByRequiredFields(entityConfig.getKey());
}
//update entityType to reflect config file.
entityType = entityTypeService_.save(entityType);
for(String permissionName: entityConfig.getPermissions()){
if(!entityTypeService_.hasPermission(entityType, permissionName)){
Permission permission = permissionSetup.getPermission(permissionName);
if(permission!=null)
//fails on below lines
permissionService._.addPermission(entityType, permission);
}
}
}
}
both the entityTypeService and the permissionService extend the above abstract class and use the same save method without alteration, addPermissions is a forloop that calls save on each permission.
The entityTypeService works, but the permissionService fails. When The permission service is called if I do em.isTransactionalEntity it returns false.
All #transactional annotations are using the spring annotation, not the javax one.
Actually, it seems as if a few of the permissions would save and others wouldn't, almost as if it's non-deterministic, but this may simple be due to modifying a database file that had some of the values already set and thus didn't need to run some of the logic the first time through.
I've done quite a bit of stumbling around but am no closer to determining what would cause my transaction to end. I had thought perhaps it was the #persistenceContext, since the JPARepos get their entityManager through a different approach then autowireing with #persistenceContext, but if that were the case everything would fail?
Any help would be appreciated, I'm pretty stumped on the cause of this.
Assuming you have enabled #EnableTransactionManagement on #Configuration class.
Since you didn't set any propagation on #Transaction the default value is Required. It means all methods must be part of transaction. Since one of your abstract methods is not part of the #Transactional hence the error.
For more information on Spring Transactions.
Note: Image taken from above link.
JBoss 4.x
EJB 3.0
I've seen code like the following (greatly abbreviated):
#Stateless
#TransactionAttribute(TransactionAttributeType.NOT_SUPPORTED)
public class EJB1 implements IEJB1
{
#EJB
private IEJB1 self;
#EJB
private IEJB2 ejb2;
#TransactionAttribute(TransactionAttributeType.REQUIRES_NEW)
public boolean someMethod1()
{
return someMethod2();
}
#TransactionAttribute(TransactionAttributeType.REQUIRES_NEW)
public boolean someMethod2()
{
return self.someMethod3();
}
#TransactionAttribute(TransactionAttributeType.REQUIRES_NEW)
public boolean someMethod3()
{
return ejb2.someMethod1();
}
}
And say EJB2 is almost an exact copy of EJB1 (same three methods), and EJB2.someMethod3() calls into EJB3.someMethod1(), which then finally in EJB3.someMethod3() writes to the DB.
This is a contrived example, but have seen similar code to the above in our codebase. The code actually works just fine.
However, it feels like terrible practice and I'm concerned about the #TransactionAttribute(TransactionAttributeType.REQUIRES_NEW) on every method that doesn't even actually perform any DB writes. Does this actually create a new transaction every single time for every method call with the result of:
new transaction
-new transaction
--new transaction
---new transaction
...(many more)
-------new transaciton (DB write)
And then unwraps at that point? Would this ever be a cause for performance concern? Additional thoughts?
Does this actually create a new transaction every single time for
every method call
No, it doesn't. The new transaction will be created only when calling method by EJB reference from another bean. Invoking method2 from method1 within the same bean won't spawn the new transaction.
See also here and here. The latter is exceptionally good article, explaining transaction management in EJB.
Edit:
Thanks #korifey for pointing out, that method2 actually calls method3 on bean reference, thus resulting in a new transaction.
It really creates new JTA transaction in every EJB and this must do a serious performance effect to read-only methods (which makes only SELECTS, not updates). Use #SUPPORTS for read-only methods
My application loads entities from a Hibernate DAO, with OpenSessionInViewFilter to allow rendering.
In some cases I want to make a minor change to a field -
Long orderId ...
link = new Link("cancel") {
#Override public void onClick() {
Order order = orderDAO.load(orderId);
order.setCancelledTime(timeSource.getCurrentTime());
};
but such a change is not persisted, as the OSIV doesn't flush.
It seems a real shame to have to call orderDOA.save(order) in these cases, but I don't want to go as far as changing the FlushMode on the OSIV.
Has anyone found any way of declaring a 'request handling' (such as onClick) as requiring a transaction?
Ideally I suppose the transaction would be started early in the request cycle, and committed by the OSIV, so that all logic and rendering would take place in same transaction.
I generally prefer to use additional 'service' layer of code that wraps basic DAO
logic and provides transactions via #Transactional. That gives me better separation of presentation vs business logic and is
easier to test.
But since you already use OSIV may be you can just put some AOP interceptor around your code
and have it do flush()?
Disclaimer : I've never actually tried this, but I think it would work. This also may be a little bit more code than you want to write. Finally, I'm assuming that your WebApplication subclasses SpringWebApplication. Are you with me so far?
The plan is to tell Spring that we want to run the statements of you onClick method in a transaction. In order to do that, we have to do three things.
Step 1 : inject the PlatformTransactionManager into your WebPage:
#SpringBean
private PlatformTransactionManager platformTransactionManager;
Step 2 : create a static TransactionDefinition in your WebPage that we will later reference:
protected static final TransactionDefinition TRANSACTION_DEFINITION;
static {
TRANSACTION_DEFINITION = new DefaultTransactionDefinition(TransactionDefinition.PROPAGATION_REQUIRES_NEW);
((DefaultTransactionDefinition) TRANSACTION_DEFINITION).setIsolationLevel(TransactionDefinition.ISOLATION_SERIALIZABLE);
}
Feel free to change the TransactionDefinition settings and/or move the definition to a shared location as appropriate. This particular definition instructs Spring to start a new transaction even if there's already one started and to use the maximum transaction isolation level.
Step 3 : add transaction management to the onClick method:
link = new Link("cancel") {
#Override
public void onClick() {
new TransactionTemplate(platformTransactionManager, TRANSACTION_DEFINITION).execute(new TransactionCallback() {
#Override
public Object doInTransaction(TransactionStatus status) {
Order order = orderDAO.load(orderId);
order.setCancelledTime(timeSource.getCurrentTime());
}
}
}
};
And that should do the trick!
I have a code that saves a bean, and updates another bean in a DB via Hibernate. It must be do in the same transaction, because if something wrong occurs (f.ex launches a Exception) rollback must be executed for the two operations.
public class BeanDao extends ManagedSession {
public Integer save(Bean bean) {
Session session = null;
try {
session = createNewSessionAndTransaction();
Integer idValoracio = (Integer) session.save(bean); // SAVE
doOtherAction(bean); // UPDATE
commitTransaction(session);
return idBean;
} catch (RuntimeException re) {
log.error("get failed", re);
if (session != null) {
rollbackTransaction(session);
}
throw re;
}
}
private void doOtherAction(Bean bean) {
Integer idOtherBean = bean.getIdOtherBean();
OtherBeanDao otherBeanDao = new OtherBeanDao();
OtherBean otherBean = otherBeanDao.findById(idOtherBean);
.
. (doing operations)
.
otherBeanDao.attachDirty(otherBean)
}
}
The problem is:
In case that
session.save(bean)
launches an error, then I get AssertionFailure, because the function doOtherAction (that is used in other parts of the project) uses session after a Exception is thrown.
The first thing I thought were extract the code of the function doOtherAction, but then I have the same code duplicate, and not seems the best practice to do it.
What is the best way to refactor this?
It's a common practice to manage transactions at one level above DAOs, in services or other business logic classes. That way you can, based on the business/service logic, in one case do two DAO operations in one transaction and, in another case, do them in separate transactions.
I'm a huge fan of Declarative Transaction Management. If you can spare the time to get it working (piece of cake with an Application Server such as GlassFish or JBoss, and easy with Spring). If you annotate your business method with #TransactionAttribute(REQUIRED) (it can even be set to be done as default) and it calls the two DAO methods you will get exactly what you want: everything gets committed at once or rolled back over an Exception.
This solution is about as loosely coupled as it gets.
The others are correct in that they take in to account what are common practice currently.
But that doesn't really help you with your current practice.
What you should do is create two new DAO methods. Such as CreateGlobalSession and CommitGlobalSession.
What these do is the same thing as your current create and commit routines.
The difference is that they set a "global" session variable (most likely best done with a ThreadLocal). Then you change the current routines so that they check if this global session already exists. If your create detects the global session, then simply return it. If your commit detects the global session, then it does nothing.
Now when you want to use it you do this:
try {
dao.createGlobalSession();
beanA.save();
beanb.save();
Dao.commitGlobalSession();
} finally {
dao.rollbackGlobalSession();
}
Make sure you wrap the process in a try block so that you can reset your global session if there's an error.
While the other techniques are considered best practice and ideally you could one day evolve to something like that, this will get you over the hump with little more than 3 new methods and changing two existing methods. After that the rest of your code stays the same.