On Thu, 19 Aug 2010, Roman Chyla wrote:
> Index size: 1.1 GB (not optimized -- therefore bigger)

Does `all' emulate our `any field' index?  Seems to have more terms, so
seems to include more fields.

It would be good to create a Lucene equivalent of some of our indexes,
say title:

   title: 210__a 245__a 246__a 242__a

and compare only that regarding number of terms, size, etc.

Also, the indexing times are good, what about the search times, say
`model NOT x-ray' in `title'?  It would be good to sit together to play
with the instance live...

Best regards
-- 
Tibor Simko

Reply via email to