Mailing-List: contact general-help@lucene.apache.org; run by ezmlm
Precedence: bulk
Reply-To: general@lucene.apache.org
Received-SPF: neutral (athena.apache.org: local policy)
From: "Uwe Schindler" <uwe@thetaphi.de>
To: <general@lucene.apache.org>
References: <24664839.post@talk.nabble.com> <4A6CA0C5.2070209@ice-sa.com>
 <24670355.post@talk.nabble.com> <p0624080fc69277056b73@[192.168.1.43]>
 <24671050.post@talk.nabble.com>
 <D4FBE517-FAFF-4220-A8AD-7CE72C4449A1@apache.org>
 <24701672.post@talk.nabble.com> <44FC07357EE24E8DB856430E5829A63C@VEGA>
 <24702697.post@talk.nabble.com>
Subject: RE: Boolean query with 50,000 clauses! Possible? Scalable?
Date: Tue, 28 Jul 2009 18:17:49 +0200
Message-ID: <0BAE8DA72BE44EF28F500743766342BC@VEGA>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
In-Reply-To: <24702697.post@talk.nabble.com>
Thread-Index: AcoPnfUbBy4gyAFQTNuCL1ApxacgNwAAFa0w

Hi Dado,

> However, that you are suggesting is to query the author index (for example
> for authors weighing more than 200lbs) and then for each author query the
> book index with the author name. The problem is: the author query could
> return tens if not hundreds of thousands author names. It would be
> unreasonable to loop through each one of them and start a query to the
> book
> index at each step.

You misunderstood me here. I meant an index for authors, if the search
results the use should get displayed are authors and the user (may) then
click on one author and display all books.
In your case I would add this weight value also to the books index: e.g. in
field name "author-weight" (and use the new NumericField from Lucene 2.9 for
it, this value is then also denormalized). Author names are in "author-name"
and so on. And then a simple NumericRangeQuery on this field.

I was wondering whether solr/lucene allows for this
> sort
> of intersections to be done at the server level, not at the client level.
> What I would like to see in lucene/solr is to query an index with an a
> large
> array of values for a specific field (in this case something like
> "book.author_name IN ([author0.name, author1.name, ..., authorN.name])"
> the
> author name array being the result of a "subquery/nested query" that could
> return a large number of hits.

Subqueries or joins are not possible per se. You should always denormalize
everything.

> Thanx for your help and consideration,
> 
> Dado
> --
> View this message in context: http://www.nabble.com/Boolean-query-with-
> 50%2C000-clauses%21-Possible--Scalable--tp24664839p24702697.html
> Sent from the Lucene - General mailing list archive at Nabble.com.