openjpa-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Patrick Linskey" <plins...@gmail.com>
Subject Re: Performance: single vs multiple transactions
Date Thu, 16 Aug 2007 13:30:55 GMT
A couple of notes / questions:

1. If you use em.delete(em.getReference(cls, id)), then you can avoid
actually loading the instances into memory, which is nice if you know
that you don't need to look at the data, but just want to delete the
records.

2. Depending on your database, you might get some benefit by setting
savepoints instead of just flushing.

-Patrick

On 8/16/07, Christiaan <christiaan_db@hotmail.com> wrote:
>
> Hi Christian,
> thanks for the reply. I am aware of the impact the combining of actions and
> flushing has on the database server performance. However, the difference
> soon (after a couple of hundred deletes) becomes more than a factor of 20!
> That seems a bit much to me. But anyway, I will research it a little more to
> see if I can find the performance bottleneck.
>
> Kind regards,
> Christiaan
> --
> View this message in context: http://www.nabble.com/Performance%3A-single-vs-multiple-transactions-tf4272115.html#a12179269
> Sent from the OpenJPA Users mailing list archive at Nabble.com.
>
>


-- 
Patrick Linskey
202 669 5907

Mime
View raw message