lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Grant Ingersoll <gsing...@apache.org>
Subject Re: Boost Single Values in Field
Date Thu, 21 Feb 2008 13:45:25 GMT

On Feb 21, 2008, at 6:53 AM, JensBurkhardt wrote:
>
> Hello again,
>
> Thanks for your immediate response. As i understand, the only way to  
> boost
> the values is on query side, right?
> The Problem is, that i need to boost field values. Okay, some more  
> details:
> I need it for a library. The string in the field i'm working with is
> something like "Title author year_of_publication some_catchwords"  
> and so on.
> It's like on big String.
> This field is called "Free-Text" so the query can something like  
> "author AND
> title".
> And that's the point, the order of the search terms is arbitrary so  
> i don't
> know what i do boost in the field. If i want to boost the title i  
> need to
> boost the specific value in the Free-Text field.
> If I'm wrong and i can handle this problem with using payload, could  
> someone
> please explain it cause i don't understand how to use payload on  
> index side.
>

I would split the string (either ahead of time or by writing your own  
TeeTokenFilter and SinkTokenizer) such that you can separate out your  
one big String into two smaller fields.  Other option, is to use  
payloads

Token.setPayload() is how you set payloads.  There are a couple  
examples in 2.3 in the contrib/analysis package of how to add payloads  
for specific things, you would just need your own that recognized your  
structure.

Also have a look at Michael B's excellent ApacheCon slides: http://people.apache.org/~buschmi/apachecon/AdvancedIndexingLuceneAtlanta07.ppt

---------------------------------------------------------------------
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