lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Aphinyanaphongs, Yindalon" <ping.p...@Vanderbilt.Edu>
Subject RE: Are similarity scores computed when using sort?
Date Sat, 27 Nov 2004 18:01:41 GMT
Thanks for the quick response.  I don't necessarily need to calculated the similarity score.
It was my understanding that the inverted index had a list of all the hits for each term so
technically, no document returned from the inverted index should have a score of 0.  Thus,
if I have some numerical field to sort by, the inverted index itself handles returning the
documents with the term.
Would you know in what java class the call is made to Similarity and where that code may be
commented out, or would it be preferable to write my own similarity sub-class instead?


From: Erik Hatcher []
Sent: Sat 11/27/2004 4:43 AM
To: Lucene Users List
Subject: Re: Are similarity scores computed when using sort?

Yes, Similarity is still computed.  It has to be in order to determine
if the documents considered are a hit or not.  Scores of 0 are not a

You certainly could simplify the Similarity computations though, by
creating your own implementation and returning 1 from all the methods.


On Nov 27, 2004, at 2:46 AM, Aphinyanaphongs, Yindalon wrote:

> I have an search application that is very performance conscious.  I've
> looked through the IndexSearcher code, and haven't been able to
> clarify whether a similarity score is calculated if the results are
> sorted by some numerical field value? Basically, it would be
> preferable to not incur the computational cost of generating a
> similarity score if it is never used.
> Thanks
> Yin

To unsubscribe, e-mail:
For additional commands, e-mail:

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message