Hi,
especially for paginating we store the results on the file system. Also, in
our case we crop the result at Nmax because usually people don't access really
all results.
Since our highlighting is based on the result of the tokenizer, we also store
the tokens right with the document in the database. The Lucene index itself
only contains the document IDs from the database, nothing else if not
absolutely necessary.
Using xdebug/wincachegrind, the part of ZSL which seeks/reads from the index
seemed to be one of the possible bottlenecks, however I don't have any numbers
with me.
- Markus
Eric Marden wrote:
Can you cache Lucene Search Results? Is there another way to speed it up?
--
Eric Marden
Sr. PHP Developer
*Bonnier Corporation*
460 N. Orlando Avenue, Suite 200
Winter Park, Florida 32789
www.bonniercorp.com <http://www.bonniercorp.com/>
