I am trying to use a join to avoid an extra select when I retrieve an object
that has a property which is another object (not a primitive or native
type).
The relevant fragments of my XML file are:
<resultMap id="YMap" class="Y">
<result property="ID" column="Y_ID"/>
<result property="Name" column="Y_Name"/>
<result property="Description" column="Y_Desc"/>
</resultMap>
<resultMap id="XMap" class="X">
<result property="ID" column="X_ID"/>
<result property="Name" column="X_Name"/>
<result property="Description" column="X_Desc"/>
<result property="MyY" resultMapping="MyStatements.YMap"/>
</resultMap>
...
<select id="X.Get" resultMap="XMap">
select * from X, Y where X.X_ID = #value# and X.Y_ID = Y.Y_ID
</select>
This works fine as long as X.Y_ID is non null. However, the database schema
allows X.Y_ID to be null. From an application perspective, I'd like to
reconstitute an instance of class X where it's MyY property (of type Y) will
be null when X.Y_ID is null.
So the above does not work. The only alternatives I can currently think of
are:
1) replace
<result property="MyY" resultMapping="MyStatements.YMap"/>
with
<result property="MyY" select="Get.Y"/>
and change X.Get to remove the join...
but that means every X.Get will invoke a Y.Get and any X.GetAll (assuming I
create one) will invoke multiple Y.Get (N+1) selects problem.
or 2)
Use caching. This is fine and I intend to at some point, but would like to
get everything working first before going the caching route.
Any ideas?
