using lucene with a very large index

[tal blum <th...@hotmail.com>]

Hi, I'm building a very large index, that contains several categories.
I have several questions I hope you can answare.
1) Is there a way to use lucene with several indexes without merging them?
2) Does the Document id changes after merging indexes adding or deleting documents?
3) Has anyone implemented a GUI to the lucene index, such that enables to deletions by id or sql-like queries?
4) assuming I have a term query that has a large number of hits say 10 millions, is there a way to get the say the top  10 results without going through all the hits?

thx tal.

Re: using lucene with a very large index

[tal blum <th...@hotmail.com>]

 > 4) assuming I have a term query that has a large number of hits say
 > 10 millions, is there a way to get the say the top  10 results
 > without going through all the hits?

 See the Javadocs for Searcher and IndexSearcher, I think you'll find
 the answer there.

 thx Otis,
but I still don't understand , because the documents are stored per Term
sorted by docId,
in order to get the top ranking document for a TermQuery you have to go over
all of the TermDocs
for that Term, this causes problems for search engine that contain many
Documents.
one solution to that is to change the implementation and store the docs
sorted by their term score.
what do you think?

tal.

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>

Re: using lucene with a very large index

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

--- tal blum <th...@hotmail.com> wrote:
> Hi, I'm building a very large index, that contains several
> categories.
> I have several questions I hope you can answare.
> 1) Is there a way to use lucene with several indexes without merging
> them?

Look at MultiSearcher class.

> 2) Does the Document id changes after merging indexes adding or
> deleting documents?

Not sure.

> 3) Has anyone implemented a GUI to the lucene index, such that
> enables to deletions by id or sql-like queries?

I haven't seen anything like it.

> 4) assuming I have a term query that has a large number of hits say
> 10 millions, is there a way to get the say the top  10 results
> without going through all the hits?

See the Javadocs for Searcher and IndexSearcher, I think you'll find
the answer there.

Otis


__________________________________________________
Do You Yahoo!?
Send FREE Valentine eCards with Yahoo! Greetings!
http://greetings.yahoo.com

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>