Re: Multiple-field queries

[Otis Gospodnetic <ot...@yahoo.com>]

Redirecting to java-user.
I wouldn't worry too much about complexity of option 1, unless you know that's just the nature of your app.
Note that going with option one you benefit from the ability to assign different boosts to different fields, length norm comes into play when scoring, etc.

Otis

----- Original Message ----
From: Enzo Michelangeli <en...@gmail.com>
To: java-dev@lucene.apache.org
Sent: Sunday, October 15, 2006 12:14:51 AM
Subject: Multiple-field queries

I need to be able to search over multiple fields at the same time. The 
suggestions I have seen so far to solve this problem are either:

1. Use a boolean query, or
2. During indexing, create an extra "all" field and add it every time 
another field is added to the document.

However, 1. may make the query very large and inefficient, and 2. is 
problematic with fields built from a Reader, because the Reader gets closed 
after the first field is indexed (and using a cloned Reader would require 
running a text extractor twice, which sounds wasteful of CPU cycles).

Would it be possible to add to the API some method to add in on stroke 
multiple fields that share a same Reader, while using that Reader only once, 
or to define fields with an array of names (such as Field f = new 
Field(String[], Reader) ) rather than a single name?

If not, are there alternative ways to achieve my goal, other than 
pre-reading the data into a String, which is obviously dangerous because it 
can cause a memory allocation failure if the end of file doesn't arrive soon 
enough?

Thanks --

Enzo


---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-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