lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Micah Jaffe <mi...@affinitycircles.com>
Subject Understanding/controlling role of Weight in IndexSearcher
Date Wed, 10 Sep 2008 03:57:10 GMT
Quick summary of situation (using 2.3.2, StandardAnalyzer):

I've taken a field that was being created as a "default" for a  
document, e.g. a giant string of glommed on values from other field  
values and instead created a boolean query to hit all the fields which  
would normally contribute to this default field.  The problem being  
that now it appears the searcher will not return documents that do in  
fact match on one of the many OR'd fields.  I was able to sleuth this  
out thanks to the little checkbox in Luke that forced it to return any  
matches on a search, but I'm a bit puzzled on how best to approach a  
fix.  I'm guessing by setting a low weight the searcher would return  
expected matches, but I'm also curious how weights are calculated.   
For instance when searching on an address field, e.g. "1234 William  
Way", the searcher wouldn't normally return a document when querying  
on "william", but an exact match on the first name field would (which  
is also boosted higher in the index, which is another thing I'm also  
wondering that might be contributing to the less-than-expected results).

thoughts?  pointers?  best practices?

thanks,
Micah

---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org


Mime
View raw message