Robert, excellent, I didn't know of the omitNorms functionality before, my
schema is already set up like this.   I think I still need to use my custom
scorer though because I've also modified the idf score.
Thanks,
Jason

On Tue, Oct 4, 2011 at 9:31 AM, Robert Muir <rcm...@gmail.com> wrote:

> On Tue, Oct 4, 2011 at 9:28 AM, Jason Toy <jason...@gmail.com> wrote:
> > Robert,
> > I think we might have a miscommunication.  The docs that require custom
> > sorting , of TYPE_ID  = 1, never sort against other types of docs.  They
> are
> > always only sorted between other docs of TYPE_ID = 1.  I still need
> custom
> > scoring because for this particular doc type though to get the docs in
> > "proper" order  I just want fieldNorm to be 1.    Since I am not sorting
> > between  the different types,  does this seem more palatable and if so
> how
> > would you do this? I see that there is/used to be queryNorm method, but
> I'm
> > unable to override it in DefaultSimilarity subclass.
> > Jason
>
> this is exactly why i said: get rid of your type_id field, create
> field1 and field2, putting the content in the appropriate field
> depending upon its type, and omit norms on field1. then you are done.
>
> instead of querying on "field:XXXX AND typeID=1" just query on
> "field1:XXXX"
>
> --
> lucidimagination.com
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
> For additional commands, e-mail: dev-h...@lucene.apache.org
>
>


-- 
- sent from my mobile
6176064373

Reply via email to