On 1/11/07, Bryan Noll <[EMAIL PROTECTED]> wrote:
Andre...
Thanks for pointing this out. Although, in this case, this is a limitation
I'm satisfied with. I personally don't think it is a good idea to use a
primitive as an identifier value. The perfect example of why to me is that
primary key columns are almost always not nullable, and must be unique by
definition. It is evident then (assuming your using a non-primitive type
for an Id) that an entity whose id value is null is transient. This is not
obvious if its value is zero? Although unlikely, who's to say that 0 isn't
the actual unique id value from the database. I know the JPA/EJB3
specification allows primitives for Id's, but I would rather intentionally
'not support' them (and document that of course) in appfuse as I feel it is
a best practice to avoid their use.
I'm cc'ing the dev list on this one. Dev folks, I would appreciate
feedback on this concept. If you disagree with my view here, please speak
up.
I big read warning on the JPA tutorial that says "don't use primitives
for primary keys" is fine with me.
Matt
André Faria wrote:
Hi All,
There is a problem in DaoJpa if you are using a primitive number (int,
double, float, long) @Id the method save always will call merge because the
objId value will return 0, not null...
That's the solution that I finded
Object objId = DaoUtils.getPersistentId(o);
if (objId == null
|| (objId instanceof Integer && objId.equals(0))
|| (objId instanceof Long && objId.equals(0L))
|| (objId instanceof Float && objId.equals(0F))
|| (objId instanceof Double && objId.equals(0D))) {
this.entityManager.persist(o);
} else {
this.entityManager.merge(o);
}
Regards,
André Faria
---------------------------------------------------------------------
To unsubscribe, e-mail:
[EMAIL PROTECTED] For additional
commands, e-mail: [EMAIL PROTECTED]
--
http://raibledesigns.com
