Hi,

Tables without primary keys is one of the features of JDO that was not adopted by JPA.

Maybe you should look at JDO implementations.

Craig

On May 20, 2009, at 8:12 AM, is_maximum wrote:


Hello

To Andrei I want to say that because it is simple to create an object and
send it to be persisted however this could be a good idea.

And to Kevin, if in secondary table we have only a foreign key to distinct records that would be enough because the id for the secondary table is not used anywhere. All we need from these two tables is a report that tells us what kind of events (secondary table) have been occurred for a specific operation in specific time (master table) and order is not important since the time of event will order those records in right way. So what is this ID good for? In case of existence of this ID we have to either create an Oracle
sequence (which is a bottleneck in database because it takes up cache
particularly in a clustered environment) or selecting the maximum ID (which is required to scan all the records) or create a sequence table managed by ORM (That will end up with a select statement and an update following that)
or create that id manually (that it is almost impossible)


Kevin Sutter wrote:

Hi is_maximum,
I'm still a little confused by your scenario. Following your described scenario... Your master table would have an Id field, but your secondary
table would not have an explicit Id field.  The foreign key from your
master
to secondary would just be some arbitrary column from the secondary table?
Do I have that right?  And, why would removing an Id field help with
performance?  You mention to get rid of its sequence, but there's no
requirement to define an Id field with a sequence.

Even though I'm still a little confused by your scenario, there are a
couple
of items to be aware of from an OpenJPA perspective.  The JPA spec
requires
an Id field, but OpenJPA does not require one.  Well, not exactly.
Instead
of declaring an explicit Id field, you could instead declare an Id via the @DataStoreId annotation [1]. This hides the Id field from your Entity
definition, but under the covers we still use an implicit Id field in
order
to insert and find records.

Another possibility is coming with the updated JPA 2 specification and the use of derived identities. I'm probably stretching this one a bit, but
this
support would allow you to derive an identity for an Entity based on the identity of a "parent" Entity. This is normally used when the dependent Entity is the owner of a many-to-one or one-to-one relationship to the parent entity. Here again, OpenJPA provides a similar functionality with
their Entities as Identity fields.

So, bottom line is that some type of Identity is required for proper
Entity
definition and usage.  But, OpenJPA (and eventually the JPA 2 spec)
provides
for some methods to get around the explicit definition of an Id field.

Hope this helps,
Kevin

[1]
http://openjpa.apache.org/builds/latest/docs/manual/manual.html#ref_guide_pc_oid_datastore
[2]
http://openjpa.apache.org/builds/latest/docs/manual/manual.html#ref_guide_pc_oid_entitypk

On Wed, May 20, 2009 at 7:14 AM, is_maximum <[email protected]> wrote:


Hi
We have some tables in which id is not important and actually it is
useless.
For example we have two logging tables, one is master and the other is keeping details. The only foreign key from the master table is enough and the second table has no relationship with other tables so if we remove
its
ID we can get rid of its sequence and this will be great in terms of
performance since this table is considered to keep lots of records and
its
purpose is for preserving events took place in the system

Now my question is how to remove its @Id from the entity since the
OpenJPA
complains if the entity has no field marked as id

Thanks
--
View this message in context:
http://n2.nabble.com/Can-we-have-an-entity-with-no-%40Id--tp2945752p2945752.html
Sent from the OpenJPA Users mailing list archive at Nabble.com.






-----
--
Regards
Mohammad
http://pixelshot.wordpress.com Pixelshot
--
View this message in context: 
http://n2.nabble.com/Can-we-have-an-entity-with-no-%40Id--tp2945752p2946693.html
Sent from the OpenJPA Users mailing list archive at Nabble.com.


Craig L Russell
Architect, Sun Java Enterprise System http://db.apache.org/jdo
408 276-5638 mailto:[email protected]
P.S. A good JDO? O, Gasp!

Attachment: smime.p7s
Description: S/MIME cryptographic signature

Reply via email to