Query to return shared data for performance reasons.

[Gustav Trede <gu...@telia.com>]

hello,

Im interested in helping openjpa to implement that query can return 
shared objects from the cache.

On many usage cases its only of interest to read data, no need to get a 
clone/copy then.
(and yes its up to the user to not actually edit the returned data.)

This exists in eclipselink and works good, QueryHints.READ_ONLY.

Could i please get some initial help,  where to do this in the code base 
in a nice way ?.

regards
gustav 

Re: Query to return shared data for performance reasons.

[Gustav Trede <gu...@telia.com>]

hello,

The situation im talking about is the current enforced cloning of the 
return data from the cache .
Its not needed to clone data in read-only situations.

This functionality is available in eclipselink.
A query hint QueryHints.READ_ONLY is used to enable return of data from 
cache "by reference",ie not a copy / cloned.


Jeremy Bauer skrev:
> Hi Gustav,
>
> This may not be exactly what you have in mind, but OpenJPA has a data
> cache for caching entities and a query result cache for caching
> certain types of query results.  Properly tuned, these caches can
> dramatically improve the performance of cache-suitable applications.
>
> Here is a link to the caching documentation.
> http://openjpa.apache.org/builds/1.1.0/docs/ref_guide_caching.html
>
> -Jeremy
>
> On Thu, Jul 3, 2008 at 10:31 AM, Gustav Trede
> <gu...@telia.com> wrote:
>   
>> hello,
>>
>> Im interested in helping openjpa to implement that query can return shared
>> objects from the cache.
>>
>> On many usage cases its only of interest to read data, no need to get a
>> clone/copy then.
>> (and yes its up to the user to not actually edit the returned data.)
>>
>> This exists in eclipselink and works good, QueryHints.READ_ONLY.
>>
>> Could i please get some initial help,  where to do this in the code base in
>> a nice way ?.
>>
>> regards
>> gustav
>>
>>     
>
>   

Re: Query to return shared data for performance reasons.

[Jeremy Bauer <te...@gmail.com>]

Hi Gustav,

This may not be exactly what you have in mind, but OpenJPA has a data
cache for caching entities and a query result cache for caching
certain types of query results.  Properly tuned, these caches can
dramatically improve the performance of cache-suitable applications.

Here is a link to the caching documentation.
http://openjpa.apache.org/builds/1.1.0/docs/ref_guide_caching.html

-Jeremy

On Thu, Jul 3, 2008 at 10:31 AM, Gustav Trede
<gu...@telia.com> wrote:
> hello,
>
> Im interested in helping openjpa to implement that query can return shared
> objects from the cache.
>
> On many usage cases its only of interest to read data, no need to get a
> clone/copy then.
> (and yes its up to the user to not actually edit the returned data.)
>
> This exists in eclipselink and works good, QueryHints.READ_ONLY.
>
> Could i please get some initial help,  where to do this in the code base in
> a nice way ?.
>
> regards
> gustav
>