You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@jackrabbit.apache.org by Paco Avila <pa...@git.es> on 2007/02/16 10:10:07 UTC
question about score
I think that the score system is buggy or I don't know how it works...
When I make this search:
//element(*, okm:document)[@okm:author='pavila']
Not all returned results have the score = 1000. Why??
--
GIT Consultors S.L.
c\ Francesc Rover 2-B
07003 Palma de Mallorca
(Illes Balears)
Re: question about score
Posted by Marcel Reutegger <ma...@gmx.net>.
Paco Avila wrote:
> All the returned nodes have the same number of properties. But the each
> one have a "okm:content" subnode where stores the document binary
> content. Is this binary content modifying the score?
the score is not just influenced by the number of the properties but also by
their size (i.e. the number of tokens they contain).
therefore, yes, if you have binary content that is fulltext indexed, it will
influence the score (iirc).
> This is a bit weird
> because I want to ckeck ONLY one property value :/
yes it is, but why do you bother when you just query for a property value?
regards
marcel
Re: question about score
Posted by Paco Avila <pa...@git.es>.
El vie, 16-02-2007 a las 14:19 +0100, Marcel Reutegger escribió:
> Paco Avila wrote:
> > I think that the score system is buggy or I don't know how it works...
> > When I make this search:
> >
> > //element(*, okm:document)[@okm:author='pavila']
> >
> > Not all returned results have the score = 1000. Why??
>
> The 'size' of the node is also taken into consideration when the score is
> calculated.
>
> The more content (properties) you have on a node the less relevant is a match
> for that node.
>
> For details see: http://lucene.apache.org/java/docs/scoring.html
All the returned nodes have the same number of properties. But the each
one have a "okm:content" subnode where stores the document binary
content. Is this binary content modifying the score? This is a bit weird
because I want to ckeck ONLY one property value :/
--
GIT Consultors S.L.
c\ Francesc Rover 2-B
07003 Palma de Mallorca
(Illes Balears)
Re: question about score
Posted by Marcel Reutegger <ma...@gmx.net>.
Paco Avila wrote:
> I think that the score system is buggy or I don't know how it works...
> When I make this search:
>
> //element(*, okm:document)[@okm:author='pavila']
>
> Not all returned results have the score = 1000. Why??
The 'size' of the node is also taken into consideration when the score is
calculated.
The more content (properties) you have on a node the less relevant is a match
for that node.
For details see: http://lucene.apache.org/java/docs/scoring.html
regards
marcel