lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Karl Wettin <>
Subject Re: XorReader?
Date Thu, 22 Oct 2009 21:24:46 GMT

22 okt 2009 kl. 20.00 skrev Chris Hostetter:

> : I'm thinking a decorator with deletions on top of the original  
> reader, merged
> : with the clone reader using a MultiReader. But this would still  
> require a new
> you don't really mean a clone do you? ... you should just need a very
> small index containing the new versions of the docs, in a  
> MultiReader with
> another reader that wraps the orriginal but overlays some deleted docs
> metadata.

Sorry, bad terminology. What I ment with "clone reader" is the reader  
of an index containing only the "by a user cloned and modified entries  
from the original database".

> : So far I haven't done anything but considering wether it's doable  
> or not.
> it sounds feasible .. but the real question is how many users you'll  
> be
> dealing with.  if it's a really small number you might be better off  
> just
> haveing one complete index per user, if it's really big then the  
> number of
> indexes you're managing could get problematic.

Lots of users at the same time. It should at least scale to handle a  
hundred of them at the same time. Cloning the original index and add  
the modifications sounds too expensive, but I might be wrong. I'm as  
worried about the effect it would have on query response time as the  
resources it would consume.

I know to little about the Searcher. What resources am I wasting by  
creating multiple Searchers on the same IndexReader? (Not really the  
same instance of IndexReader, there would be the deleted-decorator and  
the MultiReader too.) Can I expect some sort of latency (except for  
when retrieving data from the "per user index with modifications"- 
reader)? Or is this stuff associated with the IndexReader rather than  
the Searcher?


To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message