Craig L Russell schrieb:
This result is surprising, considering that we added verbiage to the specification to address this issue. I'll take a closer look, but my intent was not to allow this scenario, either in the datastore or in memory *after the transaction commits*.
I'm glad that this case was intended to be covered, at least with a JDOUserException being thrown, as the user could easily produce inconsistent data in the database, i.e. a one-to-one relationship degenerating into a one-to-many relationship. Maybe it would be useful if a JDOUserException was thrown when a one-to-one relationship is traversed and more than one related object is found (like with unique Queries), as currently it is pretty much unspecified what should happen in that case (unless some more "verbiage" has escaped me in the spec...)


The way the spec was changed near the end of the cycle, both sides of a relationship need to be consistent, both in the datastore and in memory after a flush operation.

This would be a candidate for a new TCK test to be released with a maintenance update of JDO.

The result you obtained should only be possible if both sides of the relationship were mapped to different columns in the datastore.


If two relationships (one on each side of an association) are mapped to the same column,
the field on only one side of the association needs to be explicitly mapped.
The field on the other side of the relationship can be mapped by using the mapped-by at-
tribute identifying the field on the side that defines the mapping. Regardless of which side
changes the relationship, flush (whether done as part of commit or explicitly by the user)
will modify the datastore to reflect the change and will update the memory model for con-
sistency. There is no further behavior implied by having both sides of the relationship map
to the same database column(s). In particular, making a change to one side of the relation-
ship does not imply any runtime behavior by the JDO implementation to change the other
side of the relationship in memory prior to flush, and there is no requirement to load fields
affected by the change if they are not already loaded. This implies that if the RetainVal-
ues flag or DetachAllOnCommit is set to true, and the relationship field is loaded, then
the implementation will change the field on the other side so it is visible after transaction
Conflicting changes to relationships cause a JDOUserException to be thrown at flush
time. Conflicting changes include:
•adding a related instance with a single-valued mapped-by relationship field to
more than one one-to-many collection relationship
•setting both sides of a one-to-one relationship such that they do not refer to each

On May 17, 2006, at 12:07 PM, Michael Bouschen wrote:

Hi Jörg,

you can create an unique index on the FK in the database. This still does not automatically update the other side of the relationship. But it prevents storing unexpected relationship values into the database, because the transaction would fail if a2 still refers b2.

The JDO implementation might support managing bidirectional relationships in memory, but this is not mandated by the spec.

Regards Michael

Eric Samson schrieb:

/In particular, I'd expect a2 not pointing to b2 anymore, as this clearly is supposed to be is a 1:1 relationship. The implementation can determine this being a 1:1 association, as there are two corresponding fields, and consequently it should issue /

Bu default, there is no 1:1 relationships either in Java or in RDBMS.

If you want to insure this 1:1 bi-directional link you have to:

    * use some triggers in your db
    * use a JDO callback
    * do it cleanly in your set /getters in Java
    * use an ORM tool that can manage 1:1 bi-directional links

I see. Hm. Well. For this relationship, being "bidirectional" seems to be equivalent to "being 1:1". Confusing.

Anyway, thanks for your answer!


--Michael Bouschen        Tech@Spree Engineering GmbH
Tel.:++49/30/235 520-33        Buelowstr. 66           
Fax.:++49/30/2175 2012        D-10783 Berlin           

Craig Russell
Architect, Sun Java Enterprise System
408 276-5638
P.S. A good JDO? O, Gasp!