lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Richard Jones ...@last.fm>
Subject Re: Creating document fields by providing termvector directly (bypassing the analyzing/tokenizing stage)
Date Wed, 02 Nov 2005 15:50:33 GMT
Hi Erik
Our lucene-powered music search went live this week, so your search should 
work now: http://www.last.fm/explore/search.php?q=Michael+Hedges
Before we discovered lucene our search sucked *really* badly ;)

Adding multiple fields like this is similar to what i'm doing now (i am using 
whitespace analyzer), although it may remove some of the work, i'd still have 
to add in the extra steps of building strings instead of handing over a 
termvector durectly.

I guess i need to delve into the lucene code see what's going on.

Cheers,
RJ


> last.fm using Lucene, sweet!   It has caught on with quite a number
> of friends, so I tried it just yesterday and my first query for music
> like "Michael Hedges" turned up nothing, so I was bummed.   - but it
> is a very cool service.
>
> Rather than building a string to index in this manner, perhaps adding
> each integer as an individual Field with the same name, with the term
> vector enabled, and using something like the WhitespaceAnalyzer.  To
> be honest, though, I'm not sure without digging deeper whether adding
> same-named fields in this manner messes with the term vector
> capabilities.
>
>      Erik
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
> For additional commands, e-mail: java-user-help@lucene.apache.org

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