Are there any problems with the hits.length() in luc 1.4?

[Marcus Falck <ma...@observer.se>]

Hi,

 

I'm indexing around 200 million articles in lucene. I have for the
moment added around 600000 articles. 

 

Using this technique:

5000 docs in RAMDir

Flush RAMDir to FSDir to create segmentfile with 5000.

Mergefactor 10.

 

I'm searching using the multisearcher.

 

When I had around 490 000 docs all new documents appeared to be in the
index but the count for the word "kan" was just around 8000 hits.

When the large merge occurred at 500.000 docs the "kan" count increased
to 65000 hits.

 

Any ideas`?

 

/

Regards

Marcus 

 

Re: Are there any problems with the hits.length() in luc 1.4?

["Aleksander M. Stensby" <al...@integrasco.no>]

Maybe you didn't close / open your writer/reader and that you where  
searching with a snapshot of the old index. And that after you had closed  
the writer and opened a new reader you got the correct result..?

On Fri, 30 Jun 2006 14:22:40 +0200, Marcus Falck  
<ma...@observer.se> wrote:

> Hi,
>
>
> I'm indexing around 200 million articles in lucene. I have for the
> moment added around 600000 articles.
>
>
> Using this technique:
>
> 5000 docs in RAMDir
>
> Flush RAMDir to FSDir to create segmentfile with 5000.
>
> Mergefactor 10.
>
>
> I'm searching using the multisearcher.
>
>
> When I had around 490 000 docs all new documents appeared to be in the
> index but the count for the word "kan" was just around 8000 hits.
>
> When the large merge occurred at 500.000 docs the "kan" count increased
> to 65000 hits.
>
>
> Any ideas`?
>
>
> /
>
> Regards
>
> Marcus
>
>



-- 
Aleksander M. Stensby
Software Developer
Integrasco A/S
aleksander.stensby@integrasco.no
Tlf.: +47 41 22 82 72

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