Re: Antw.: [eclipselink-users] Caching read-all queries

[Tom Ware <tom.ware@xxxxxxxxxx>]

Hi Zarar,

  EclipseLink assumes multiple applications will be updating your database. 
With multiple applications, there is the possibility that rows will be added to 
tables outside of the knowledge of a single application.  As a result when you 
Read All objects, EclipseLink cannot assume it's cache knows about all the 
objects - therefore, it must go to the DB to retrieve the IDs of the objects 
that the query should return.  After the IDs are returned the cache is used for 
optimization.

  If your application is designed in such a way that you know that it is the 
only application that will touch the database, or that you don't care what other 
applications add, there are a number of EclipseLink features that allow you to 
do Cache-only queries.  You can run queries that check the cache only, and you 
can run queries that cache their results and always return those results.  The 
question is, is that really what you want?

-Tom

Zarar Siddiqi wrote:
> Hi,
>
> So queries like returning a list of states or last year's items sold
> will always result in a query being executed? I find it really hard to
> believe that it's a feature that is lacking.
>
> Thanks
> Zarar
>
> On Thursday, May 20, 2010, JAAP SPIERING <jaap.spiering@xxxxxxxxxx> wrote:
> > Hi Zahar,
> > I believe this question popped up on this list a couple of days ago as well, but I couldn't find it anymore.Anyway, ReadAllQueries always hit the database, because they need to determine what records are to be returned (there might be different ones in the database compared to when your query was last executed). When records are returned that are already stored in the cache, it will return the corresponding cached objects so there is some optimization there.
> > regards,
> > Jaap
> > ----- Oorspronkelijk bericht -----
> > Van: zarars@xxxxxxxxx
> > Aan: eclipselink-users@xxxxxxxxxxx
> > Verzonden: Donderdag 20 mei 2010 00:11:39 GMT +01:00 Amsterdam / Berlijn / Bern / Rome / St!
> >  ockholm / Wenen
> > Onderwerp: [eclipselink-users] Caching read-all queries
> >
> > I'm using 1.2 and have a couple cache related questions.
> >
> > Do read-all queries (ones using getResultList()) always hit the database?
> >
> > As I understand the docs, the "CheckCacheThenDatabase" hit only
> > applies to read-object queries (ones using getSingleResult()).  In my
> > app, I have a ton of queries that are candidates for caching, but they
> > all use getResultList(). I've set an @Cache(expiry=60000) annotation
> > on those entities and when I execute them, I keep seeing the queries
> > printed in the logs.
> >
> > To me, this isn't the greatest behaviour since I'm executing the exact
> > same query with the exact same parameters.  Shouldn't EclipseLink have
> > a way of optimizing this?
> >
> > Thanks,
> > Zarar
> > _______________________________________________
> > eclipselink-users mailing list
> > eclipselink-users@xxxxxxxxxxx
> > https://dev.eclipse.org/mailman/listinfo/eclip!
> >  selink-users