Eclipse Community Forums: EclipseLink

Home » Eclipse Projects » EclipseLink » EclipseLinkException(Explanation on the pattern of exceptions in javax.persistence)

EclipseLinkException [message #515675]

Fri, 19 February 2010 12:53

Eclipse User

EclipseLinkException is the root RuntimeException in the package org.eclipse.persistence.exceptions. This is similar to the manner of PersistenceException in the package javax.persistence. My question is why doesn't EclipseLinkException actually extend javax.persistence.PersistenceException?

In my application I programmed to JPA as specified in Java EE, but when I use the JPA implementation I don't get exceptions that adhere to the specification, they only mimic it. In theory, my application could swap JPA implementations and still work if my application only exposed itself to the Java EE specification.

Specifically, I'm catching the PersistenceException in my source, but the library is throwing EclipseLinkException, so I don't catch it and miss the opportunity to rollback my transaction. The fix is to step up one level and just catch java.lang.RuntimeException, but that adds more generality when I'm only dealing with persistence issues. Otherwise I can catch EclipseLinkException directly and lose some portability. The question isn't critical, but I am curious if someone has perspective on how API implementations map to the API specification.

Re: EclipseLinkException [message #515710 is a reply to message #515675]

Fri, 19 February 2010 16:06

Eclipse User

Hello,

PersistenceException is thrown when using JPA, EclipseLinkExceptions are thrown when using EclipseLink api directly (ie session.executeQuery(query) etc). EclipseLinkException does not extend PersistenceException to avoid having a JPA dependency within the native EclipseLink api.

If your application is using JPA only, you should only be getting javax.persistence.PersistenceExceptions. Can you provide an example where you see an EclipseLinkException from JPA methods?

Best Regards,
Chris

[Updated on: Fri, 19 February 2010 16:08] by Moderator

Re: EclipseLinkException [message #515719 is a reply to message #515710]

Fri, 19 February 2010 17:19

Eclipse User

Thanks Chris,

I used some Entity classes to create an H2 database that were originally developed for Oracle database. Some of the entities Ids used sequences. I let the API recreate the database based on the entities. After I started my application I created new objects for those entities and when I committed the transaction I got the EclipseLinkException because the sequences didn't exist in the new H2 database. Obviously, not creating the sequences was the underlying problem, but the application I wrote didn't handle the exception so my transaction wasn't closed. After I added the sequences from another connection my application state was out of whack and when I tried to begin a new transaction the last one was still open, so I'd get the appropriate IllegalStateException. Restarting the application returned it to a healthy state, but I was curious how that exception escaped my grasp.

If I understand your response correctly, the nonexistent sequences lie outside the scope of a JPA exception, so that's why I got the EclipseLikeException. Or is that any transaction against the database exposes you to underlying JPA implementation?

If you'd still like an example I can provide one.

Re: EclipseLinkException [message #517933 is a reply to message #515719]

Tue, 02 March 2010 09:53

Eclipse User

Could you please provide the exception stack trace.

Previous Topic:	Adapting to existing database
Next Topic:	Question regarding JPA Joins and predicates

Goto Forum:

-=] Back to Top [=-

Current Time: Tue Jul 01 16:27:37 EDT 2025

.:: Contact :: Home ::.

Breadcrumbs

Sign up to our Newsletter