Don't mean to hijack the thread, but on this same subject, I noticed
myself that running a wildcard search '*' on even a very small index
(handful of docs, approx 50kb) seems to take an awfully long time and
I wondered whether that was usual?
Thanks
Nick
On 21 Oct 2008, at 14:51, Endijs Lisovskis wrote:
Thanks for your response. At least someone :)
But your suggestion will not make any difference in performance,
because, as
I said in first email index IS optimized.
Maybe there are some potential problems with index if data is in
UTF8 ?
How to trace - where to look for problem?
Index was made in this manner: select first 1000 articles -> add to
index ->
optimize, select next 1000 articles -> add to index -> optimize. Of
course:
1 article = 1 document. At the end index is optimized and consists
from
~7350 documents.
I hope someone can help.
Endijs Lisovskis
Jurriën Stutterheim wrote:
Hi,
You should probably optimize your index before searching (and make
sure it stays optimized, using a Cron-job).
See
http://framework.zend.com/manual/en/zend.search.lucene.index-creation.html#zend.search.lucene.index-creation.optimization
for more details on how to do this =)
- Jurriën
--
View this message in context:
http://www.nabble.com/Zend_Search_Lucene---how-to-find-performance-problem-source--tp20085562p20090740.html
Sent from the Zend Framework mailing list archive at Nabble.com.
