lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Erick Erickson" <erickerick...@gmail.com>
Subject Re: Architecture for indexing/searching mailing list archives
Date Mon, 24 Jul 2006 13:25:58 GMT
I suspect that 3 is actually better. You can use CachingWrapperFilter to
cache the filters automatically. Also, I found that filters were much faster
to construct than I first thought.

That said, though, why bother with a filter? Why not just make the list part
of the query and let Lucene take care of it for you? So each query would
have a clause 'And list-id = <specific list here>). There was a discussion
last week about Lucene's attempts to restrict queries by the
most-restrictive clause first, so it seems to me that this is all done for
you without filters. I'd only think about filters *after* I had some
performance numbers.....

Philosophically (don't you hate that word) I'm reluctant to get any more
complicated than I have to before I have some performance numbers indicating
that complexity is required...


Best
Erick

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message