lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Antonio Bruno <>
Subject I: Lucene id generation
Date Tue, 19 Dec 2006 18:45:35 GMT
The problem in reality consists on the fact to have an only dictionary of the terms for all
the fields. If the dictionary of the terms is the many large performances of a search they
diminish, even if the search is made on a single term. Then it would be wanted to be able
to index the fields of a single document on more various indices. In such a way when if I
want to create a filter on a field I use only a precise index that has a dictionary of the
terms whose largeness is function only of the terms of that field. In so far as I mean that
the searches would be efficient to the maximum. 
You notice yourself that the this case does not have nothing to that to see with the MultiSearcher…

I succeed to explain to me? How could be made one what of this type at the moment? 
Antonio Bruno
(T&S - Cosenza Italy)

> But you can do something very similar and very quickly using a unique ID
> (not the Lucene ID) that's shared across the indexes (assuming I'm reading
> your issue correctly). Then use TermDocs/TermEnum and create your filters
> that way.
> I predict endless problems with user (programmer) errors if Lucene allowed
> one to specify the (lucene) document ID.
> I suspect it just ain't going to happen since nobody has yet made a case for
> why this would be enough better than the current system to be worth the
> hassle.
> Best
> Erick

Do You Yahoo!?
Poco spazio e tanto spam? Yahoo! Mail ti protegge dallo spam e ti da tanto spazio gratuito
per i tuoi file e i messaggi 
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message