Hy Guys,
I intend to extend DocumentStoredFieldVisitor class like this:
class DocumentStoredNonRepeatableFieldVisitor extends
DocumentStoredFieldVisitor {
@Override
public Status needsField(FieldInfo fieldInfo) throws IOException {
return fieldsToAdd == null || fieldsToAdd.contains(fieldInfo.name)
? Status.YES
: fieldsToAdd.size() < doc.getFields().size() ? Status.NO :
Status.STOP;
}
}
The gain in our application is that we utilize the usage of Status.STOP.
We have 98 fields currently and in some cases we need loading just 1, 2
or 3 fields per matching doc and using this we could skip lot of field
extraction if happen matching fields to be firstly visited.
I do not have time to see how is applied the method doc.add(new
StoredField(fieldInfo.name, value)) in Document class - when comes field
with name already existing in the document does it change the existing
"value" for that field by concatenating two values, or just adds
somewhere the new value and concatenation is made later. In latter case
I think my suggested class could be applied only in cases when we do not
index fields with the same names to same documents, otherwise with
checking (fieldsToAdd.size() < doc.getFields().size()) we could stop
loading fields for that document and in this way miss some repeating
field. This is why my class is called
"DocumentStoredNonRepeatableFieldVisitor".
As we use Lucene not like a jar-file but the sources - I will change
visibility of DocumentStoredFieldVisitor's members "doc" and "fieldsToAdd".
I just want to suggest in the original class DocumentStoredFieldVisitor
those members's visibility to also be changed so that the class to be
extended in applications that use jar-files.
Cheers,
Ivan
---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org
For additional commands, e-mail: java-user-h...@lucene.apache.org
