This is similar to the issue in Java Persistence with Hibernate where there is a one-to-many relationship from item to bid as well as a one-to-one relationship between them on successful bid. In section 8.1.2, "Arbitrary join conditions with formulas" the single row is identified by a single property, "SUCCESSFUL".

My situation is that I also have a one-to-many relationship between two tables but I need to restrict the many to a single, currently effective row which is determined by testing "eff_ts = exp_ts" on the many side. The columns involved are timestamps.

I've tried variations on the <properties> and <formula> tags specifying the 1:m surrogate key join and sql on the two columns in the same, single row but with no luck.

The sql would be: "select b.beta_id from alpha a, beta b where a.alpha_id = b.alpha_id and b.eff_ts = b.exp_ts"

I already have the set of b instances loading into the alpha object, I just need the single, current beta object of the set as a one-to-one mapping.

In your posting it is unclear if you already have a mapping that you can't change (and therefore you are asking for code help on how to handle the problem), or you can change your mapping file and are asking for help on that.

If your mapping can't change, you can retrieve the "a" along with a collection of multiple "bs" and then apply a filter on the collection. Or you can write HQL .

If you want to change the mapping files, you can use a formula (a simple formula, don't complicate yourself with joins), or you can create a "subselect entity" for B, containing only the items that comply with that "ts" condition. The latter sounds like the most valuable solution to me, business-wise.