Hi Davide,
as Index option I would use Index.NOT_ANALYZED_NO_NORMS; ANALYZED is
not a proper fit as it might break the marker token in pieces and get
some unpredictable behaviour.
About Store .. do you need to store it? I assume you do to mark the
difference between a null collection and an empty collection, but
there might be ways around that, in fact we only want to know if this
document matches a Term (fieldName, nullToken), and you don't need to
store it for that.
I wouldn't disagree in using Store.YES if there are good reasons
though; could you explain how this is working?
Cheers,
Sanne
2011/8/12 Davide D'Alto <daltodavide(a)gmail.com>:
Hello,
I've created a prototype for HSEARCH-670:
https://github.com/DavideD/hibernate-search/commits/HSEARCH-670
It seems to work but I'm not sure about the options to use indexing
the null value.
At the moment it uses the following LucenOption object:
LuceneOptions options = new LuceneOptionsImpl(
Store.YES,
org.apache.lucene.document.Field.Index.ANALYZED,
org.apache.lucene.document.Field.TermVector.NO,
1F );
And I use the following default TwoWayStringBridge:
TwoWayStringBridge fb = new TwoWayStringBridge() {
@Override
public String objectToString(Object object) {
return nullMarker;
}
@Override
public Object stringToObject(String stringValue) {
return null;
}
};
This is the method:
https://github.com/DavideD/hibernate-search/commit/52a9280321e6ea06ff8324...
Do they make sense?
Cheers,
Davide
_______________________________________________
hibernate-dev mailing list
hibernate-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/hibernate-dev