Re: [eclipselink-users] JPA flush and identity/serial columns

[Tom Ware <tom.ware@xxxxxxxxxx>]

Hi Laird,

  For objects with database-based sequencing, where possible, Eclipselink will 
assign the identifier on the persist call.

-Tom

Laird Nelson wrote:
> If I am using GenerationType.IDENTITY, then must I flush() any time I 
> make use of the just-allocated identifier?
>
> In the case of saving an object graph, will EclipseLink do this for me?  
> I am seeing (as a counterexample) that Hibernate 3.5 won't.  That is, if 
> I persist a new object (which won't have an identifier until a trip to 
> the database has been made), and then in the same transaction set a 
> @ManyToOne relationship to point to it, should it be the case that the 
> JPA provider should take care of sequencing these database operations?  
> I would think so.
>
> We're seeing a situation which may prove to be the last straw in our 
> adoption of Hibernate that seems to indicate that this kind of 
> sequencing is not occurring.  Before I laboriously switch over to 
> EclipseLink I want to make sure this isn't a specification thing.
>
> Best,
> Laird
>
> --
> http://about.me/lairdnelson
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> eclipselink-users mailing list
> eclipselink-users@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/eclipselink-users