[hibernate-dev] Strange sequence effect with ORM 5

[Steve Ebersole]

Or, if you have *explicit* expectations you could, you know, explicitly
define them :)

On Fri, Aug 21, 2015 at 10:30 AM Steve Ebersole <steve at hibernate.org> wrote:

> Well you can do that.  Thats just not historically how Hibernate worked.
>
> Set `prefer_sequence_per_entity` to true
>
> On Fri, Aug 21, 2015 at 10:26 AM Sanne Grinovero <sanne at hibernate.org>
> wrote:
>
>> Ok, thanks and sorry I missed this :)
>> I was aware that it would switch to sequences rather than identity,
>> but I missed that there would be a single Sequence shared among
>> types... I still find that a bit surprising as I would expect to
>> generally want different sequences for each type.
>>
>> Thanks,
>> Sanne
>>
>> On 21 August 2015 at 16:21, Steve Ebersole <steve at hibernate.org> wrote:
>> > Please see the migration guide and/or blogs :)
>> >
>> > Yes, we changed the default mapping for how id generators are mapped as
>> > discussed extensively on the mail list last month or earlier this month.
>> > Essentially we changed the default for
>> `hibernate.id.new_generator_mappings`
>> > from false to true for 5.0.  For H2, the old mapping would have been an
>> > IDENTITY column.  The "new" mapping is the SequenceStyleGenerator.
>> >
>> >
>> >
>> > On Fri, Aug 21, 2015 at 10:16 AM Sanne Grinovero <sanne at hibernate.org>
>> > wrote:
>> >>
>> >> Hi all,
>> >> after upgrading Hibernate Search to Hibernate ORM 5.0.0.Final we
>> >> noticed a small change in behaviour in sequences.
>> >>
>> >> In a single test we persist two entities of differnent types, Country
>> >> and Address, and both types have the id mapped as follows:
>> >>
>> >>    @Id
>> >>    @GeneratedValue
>> >>    Long id;
>> >>
>> >> Previously, when running on H2, the IDs assigned to these two entities
>> >> would be respectively 1 and 1.
>> >> When upgrading ORM from 5.0.0.CR3 to 5.0.0.Final, the assigned ids are
>> >> now 1 and 2.
>> >>
>> >> Are sequences now being shared by default across unrelated entities?
>> >> Is this the new intended default?
>> >>
>> >> Thanks,
>> >> Sanne
>> >> _______________________________________________
>> >> hibernate-dev mailing list
>> >> hibernate-dev at lists.jboss.org
>> >> https://lists.jboss.org/mailman/listinfo/hibernate-dev
>>
>