There are a few solutions until we address this case in a more direct
way (which I am ashamed to say we do not now) -
1. For in-memory filtering, you can create a cover method "getSize()"
to work around JavaBean naming.
2. For DB query, if you don't have a lot of objects in those tables,
you can use a modified advice from here:
http://cwiki.apache.org/CAY/not-in-to-many-faq.html
SelectQuery excludedQuery = new SelectQuery(House.class);
// add joint prefetch that will result in an inner join to ensure
houses without residents are excluded
excludedQuery.addPrefetch("residents").setSemantics
(PrefetchTreeNode.JOINT_PREFETCH_SEMANTICS);
List excluded = ctxt.performQuery(excludedQuery);
List all = ctxt.performQuery(new SelectQuery(House.class));
all.removeAll(excluded);
3. The most efficient approach is SQLTemplate that fetches the right
objects on the first attempt.
4. My favorite approach - fetch residents instead of houses, and then
build a set of distinct houses by iterating via the list of residents
in memory. Again the efficiency depends on how big is the residents
list.
Andrus
On Oct 6, 2006, at 10:51 PM, Arturo Pérez wrote:
Hi all,
I thought I saw this somewhere on the wiki but I can't find it.
I have an object with a toMany relation (e.g. house.getResidents()).
How do I write an Expression to find all the houses with no residents?
I know this doesn't work:
Expression exp = ExpressionFactory.matchExp("residents", null);
and this doesn't work because the size() method doesn't follow bean
convention:
Expression exp = ExpressionFactory.matchExp("residents.size", new
Integer(0));
I would be much obliged...
-arturo
