lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Erick Erickson" <erickerick...@gmail.com>
Subject Re: Lucene ID and scoring.
Date Tue, 16 Oct 2007 13:35:18 GMT
Can't answer the second question, but the answer to the first is "no".
Not only are Lucene IDs internally generated, but they change when
you  delete/optimize.

Would caching filters help? Especially if you pre-computed them
at, say, warm up?

What problem are you trying to solve anyway? A statement of the
problem might get you better answers.....

Best
Erick


On 10/11/07, sandeep chawla <sand.chawla@gmail.com> wrote:
>
> Hi All,
>
> I have two questions here
>
> 1- Lucene generates id for every document it indexes and it does in
> order..0,1,2,3.....n .Can i make lucene use my id ( my id for a document
> ..say randomly generated unique id for a document) as lucene id.
>
> The reason i am asking this question because lucene filter uses lucene id
> (
> to create  BitSet ) and i dont wanna keep multiple ids .
>
> 2- Is there any way to change lucene's scoring formula without changing
> the
> lucene core.
> I know, there is one way of doing this is to extend the similarity and
> thats
> really simple but it can help only in formulae which has tf and idf in
> multiplication.
>
> for example a formula like  tf(power n) * idf ( power m) can be easily
> implemented using similarity
>
> but what about the formula tf(power n)  + idf(power m )
>
> is there a simple way to implement the above without changing lucene core
> (
> i know one way to do is too change the TermScorer and hardcore the + sign
> but that's too naive :)
>
> Thanks a lot ...
> Sandeep
>
> --
> SANDEEP CHAWLA
> House No- 23
> 10th main
> BTM 1st  Stage
> Bangalore Mobile: 91-9986150603
>

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