Hi,

I'm trying to do a unidirectional one-to-one mapping for a legacy database where the primary key of the other side of the join is available on the owning side, but not the other way around...



From the docs I get the impression this could be done with a formula:

"Almost all one to one associations map to the primary key of the owning entity. In the rare case that this is not the case, you may specify a some other column, columns or expression to join on using an SQL formula."

But I'm not clear on what the syntax would be. I've taken a look at the org.hibernate.test.onetoone.formula example, but that is a slightly different scenario.

Also, this mapping is for read-only data.

Any pointers would be most appreciated.

Thanks,

Ian

Hi,
I have set up a small example. It deals with unidirectional one-one associations from both sides, separately.
( with and without formula )

Tables:
X:
id number
name varchar2
Y:
id number
name varchar2
ref_id number ( refers to X's id )


1) One to One , unidirectional from X to Y.
<class name="X" table="x">
<id name="id" column="id" type="integer" />
<property name="name" type="string" column="name"/>
<one-to-one class="Y" name="y" property-ref="ref_id" />
</class>

<class name="cars.models.Y" table="y">
<id name="id" column="id" type="integer" />
<property name="name" type="string" column="name"/>
<property name="ref_id" column="ref_id"/>
</class>

2) One to One , unidirectional from Y to X.

i) one way.
<class name="cars.models.X" table="x">
<id name="id" column="id" type="integer" />
<property name="name" type="string" column="name"/>
</class>

<class name="Y" table="y">
<id name="id" column="id" type="integer"/>
<property name="name" type="string" column="name"/>
<one-to-one class="X" name="x" >
<formula>ref_id</formula>
</one-to-one>
</class>

ii) another way. ( X remains the same )
<class name="Y" table="y">
<id name="id" column="id" type="integer"/>
<property name="name" type="string" column="name"/>
<many-to-one class="X" name="x" column="ref_id" />
</class>

-------------------------------------------------
Rate the reply if you find it helpful

Thanks, that worked perfectly. Though I wish I understood formula a little better. So "ref_id" gets translated to "[y].ref_id=[x].[id]" or "[x].[id]=?" (where ? is the previously selected value of [y].ref_id) for joins/subselects or extra selects respectively?

Not to complain, as I'm getting a lot for free here, but is this explained somewhere outside of the source that I'm missing? If not I'd like to take a stab a contributing some documentation.

Thanks Again,

Ian

Hi Ian,
Was out of desk for some time.
>>> So "ref_id" gets translated to "[y].ref_id=[x].[id]" or "[x].[id]=?" (where ? is the previously selected value of [y].ref_id) for joins/subselects or extra selects respectively?
I don't get you completely ? Can you put more light into it ?
For the mapping shown above, we get [y].ref_id=[x].[id] in the query.