[prev in list] [next in list] [prev in thread] [next in thread] 

List:       lucene-user
Subject:    Re: Fast access to a random page of the search results.
From:       mark harwood <markharw00d () yahoo ! co ! uk>
Date:       2005-03-08 9:56:37
Message-ID: 20050308095637.10385.qmail () web26002 ! mail ! ukl ! yahoo ! com
[Download RAW message or body]

> But I suppose for Document
> has to be further subclassed so that the other
> non-initialized fields can be obtained as well, or

I don't think Document would be the right place for
this - as a design pattern it is cast as a "value
object" or "transfer object" which is passed to
(potentially remote) clients. It shouldn't be holding
connections to an index so it can load extra fields on
demand.

Regardless, any "laziness" to the selective field
loading will be hard to optimize when it is not
possible to directly access an individual field on
disk - the offsets are not stored and you have to
serially read through all of the document's fields on
disk until you reach the one you want.

There's a price to pay for allowing clients too much
freedom and I think lazy loading of field values is an
example of something which is too costly.
I personally prefer a search interface which requires
clients to state up front what fields they want
returned - the equivalent of SQL's "Select" clause in
addition to the usual "where.." matching criteria. 

Cheers,
Mark


Send instant messages to your online friends http://uk.messenger.yahoo.com 

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

[prev in list] [next in list] [prev in thread] [next in thread]