Hi Jackson, Besides the extra field being stored in your entities classes when they're serialized the only other side effects are a couple of additional methods being added to the entity when it is enhanced. As a result the serialized entity will be a little "heavier" than it used to be.
To determine the root cause of the problem one would have to look at how the application handles the entities - how they're passed from one component to another. You mentioned that you're only using local interfaces on your EJBs so as I understand it the entities shouldn't be serialized. I'm not an EJB container expert though and you might be better served to check with WebSphere support (the EJB team probably doesn't monitor this mailing list). Sorry I can't help with the local question, but I hope you at least have a work around. -Mike On Tue, Mar 11, 2008 at 8:19 AM, jackson12 <[EMAIL PROTECTED]> wrote: > > Mike, > > After setting the DetachStateField property, it does solve the issue. I am > a > little bit concerned about this fix without knowing the root cause? does > this property have any side impact? > > thanks > > > > > Michael Dick wrote: > > > > Hi Jackson12, > > > > I don't think the entities would be serialized if they're just being > > passed > > through the local interfaces on your beans. Have you tried setting the > > DetachStateField property, and verified that it does resolve the issue? > > You > > might be hitting a different scenario which has the same symptom. > > > > -Mike > > > > > > On Fri, Mar 7, 2008 at 6:09 PM, jackson12 <[EMAIL PROTECTED]> wrote: > > > >> > >> Hi Mike, > >> > >> We run into this issue. but in our case, we only use local EJBs, why > the > >> entities get serialized and deserialized? > >> > >> thanks > >> > >> > >> mikedd wrote: > >> > > >> > Hi > >> > > >> > I've seen this happen when an entity is serialized and deserialized. > >> > During > >> > the serialization process the State of the entity was lost. So when > the > >> > entity was reattached OpenJPA couldn't differentiate between an > >> unloaded > >> > field and one that was intentionally set to null. > >> > > >> > To resolve the problem you can add the following property to > >> > persistence.xml > >> > : > >> > > >> > <property name="openjpa.DetachState" > >> > value="fgs(DetachedStateField=true)" /> > >> > > >> > The real key here is the DetachedStateField=true part which instructs > >> > OpenJPA to save the State of the entity when it is serialized and > >> > deserialized. > >> > > >> > This option needs to be in your persistence.xml when the entities are > >> > enhanced. If you use static enhancement via the PCEnhancer tool > you'll > >> > have > >> > to re-run the tool after updating persistence.xml. If you don't use > >> static > >> > enhancement the container will enhance your entities for you - so > >> you'll > >> > just need to redeploy. > >> > > >> > Hope this helps, > >> > -Mike > >> > > >> > On Fri, Mar 7, 2008 at 2:28 PM, <[EMAIL PROTECTED]> wrote: > >> > > >> >> Yes, that's exactly what I am seeing, and perhaps I didn't give > enough > >> >> detail... > >> >> > >> >> In my case, the EJB service fetches the entity, it gets detached > when > >> it > >> >> moves from the EJB container to the web tier, user updates the > entity, > >> >> pass > >> >> it back to the service where it's "merged"... the fields that were > >> nulled > >> >> are not getting persisted. It has to do with detaching the entity > >> then > >> >> re-attaching it again. > >> >> > >> >> I've found a HACK (emphasis on HACK) to solve it in the interim.. I > >> >> basically take the detached entity, re-fetch the managed one, copy > >> >> properties from the detached to the managed entity, then let the > >> managed > >> >> entity persist normally. But it gets tedious when you're dealing > with > >> >> entities with 30-40 attributes and you can't use the Commons > >> >> copyProperties > >> >> because of version columns, etc. (fields that are restricted). > >> >> > >> >> If there is a better solution, I'd be more than happy to entertain > it! > >> >> > >> >> Thanks! > >> >> Randy > >> >> > > >> >> > From: Tedman Leung <[EMAIL PROTECTED]> > >> >> > Date: 2008/03/07 Fri PM 01:34:32 CST > >> >> > To: users@openjpa.apache.org > >> >> > Subject: Re: Newly null fields not being persisted? > >> >> > > >> >> > I reported this as a bug earlier and I never heard back, it still > >> >> occurs. > >> >> > A more detiled explanation of what I found is as follows : > >> >> > > >> >> > - If you have a new entity with null fields, the fields will save > as > >> >> null. > >> >> > > >> >> > - If you have an entity which is attached, you can nullify fields > as > >> >> long > >> >> > as it persists itself. i.e. without you calling the merge method. > >> >> > > >> >> > - If you retrieve an entity and detach it, then null a field, then > >> >> > reattach it (with merge), it will not nullify the field. > >> >> > > >> >> > > >> >> > if anyone knows the code well enough to fix it or point me to the > >> area > >> >> > where the logic of this is done that would be good. > >> >> > > >> >> > > >> >> > > >> >> > On Fri, Mar 07, 2008 at 01:28:40PM -0600, > >> [EMAIL PROTECTED]: > >> >> > > All, > >> >> > > > >> >> > > I'm not sure if this is a problem in MY code or OpenJPA. But, > >> here > >> >> are the symptoms in WAS w/ EJB Feature pack: > >> >> > > > >> >> > > User enters data into a form that is backed by a managed OpenJPA > >> >> entity. Entity gets persisted successfully. User goes back to the > >> same > >> >> page and nullifies one of the fields (i.e., deletes a value from one > >> of > >> >> the fields in the form). Entity gets persisted, but, the field that > >> was > >> >> nulled out is not updated. > >> >> > > > >> >> > > Looking at the generated SQL in the WAS log, I see the first > >> update > >> >> that added the data, but in the second update statement, the field > >> that > >> >> was > >> >> nulled is not included in the update (thus leaving data in the > >> underlying > >> >> table). > >> >> > > > >> >> > > Is there a setting in OpenJPA that turns this off? Or, could it > >> >> (most > >> >> likely) be a problem in my code somewhere? > >> >> > > > >> >> > > Any suggestions on how I can track this down and eliminate it? > >> >> > > > >> >> > > Cheers! > >> >> > > Randy > >> >> > > > >> >> > > >> >> > -- > >> >> > Ted > Leung > >> >> > > >> >> [EMAIL PROTECTED] > >> >> > > >> >> > You know things are getting a little fishy when you're commenting > >> out > >> >> > comments. > >> >> > > >> >> > >> >> > >> > > >> > > >> > >> -- > >> View this message in context: > >> > http://www.nabble.com/Newly-null-fields-not-being-persisted--tp15904016p15908495.html > >> Sent from the OpenJPA Users mailing list archive at Nabble.com. > >> > >> > > > > > > -- > View this message in context: > http://www.nabble.com/Newly-null-fields-not-being-persisted--tp15904016p15976501.html > Sent from the OpenJPA Users mailing list archive at Nabble.com. > >
