Hi Arthur,

Apache Lucene (Java implemenation) doesn't allow wildcard symbols at the
beginning of pattern and limits number of matched terms by 1024. It
limits search time.

I am not sure, what is preferable, to set hard limitations or give an
API to make it confugurable...


With best regards,
   Alexander Veremyev.
 

> -----Original Message-----
> From: arthur-8 [mailto:[EMAIL PROTECTED] 
> Sent: Saturday, August 04, 2007 12:22 AM
> To: fw-general@lists.zend.com
> Subject: [fw-general] Zend_Search_Lucene: Manging large 
> wildcard queries
> 
> 
> First off, a big thanks to Alexander for implementing 
> Wildcard queries in svn.
> 
> We're indexing and tokenizing quite a large amount of data 
> (mostly plaintext from .txt files with some Stopwords 
> removed). Doing a wildcard query often takes very long to 
> finish and mostly PHP just time-outs.
> 
> Currently we're doing a non-wildcard search when the search 
> string contains more than two words (which is not very 
> effective, but mostly reduces the time-outs). 
> 
> My question would be if any of you know an effective way to 
> manage Wildcard queries on many tokenized words?
> 
> Thanks.
> --
> View this message in context: 
> http://www.nabble.com/Zend_Search_Lucene%3A-Manging-large-wild
> card-queries-tf4214305s16154.html#a11989387
> Sent from the Zend Framework mailing list archive at Nabble.com.
> 
> 

Reply via email to