lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From robert engels <reng...@ix.netcom.com>
Subject Re: prefix query help ?
Date Wed, 16 Apr 2008 05:24:51 GMT
Please correct me if I am wrong, but I think the prefix query finds  
matching terms first and puts them in an array???

So if you have a numeric field, searching for 1* could give a near  
infinite number - causing an OOM.

Causes me to think though... in most cases the number of unique names  
that start with r should be relatively small - not enough to cause an  
OOM - so I need to investigate further what is happening.


On Apr 15, 2008, at 9:05 PM, Marvin Humphrey wrote:

>
> On Apr 15, 2008, at 4:07 PM, robert engels wrote:
>
>> we use this for name searching, so that you might have
>>
>> "engels r*"
>>
>> the problem being that the phrase "engels r" is quite unique, but  
>> the term r* is not unique at all.
>>
>> Anything improve the handling of this in the current lucene?
>
>
> You could tokenize differently and index pairs together.  Think  
> searching for "engles_r*".
>
> If that's not an option, make sure that those two are being joined  
> via AND -- I think that will trigger the skipTo optimization when  
> the posting lists for r* are scanned.
>
> HTH,
>
> Marvin Humphrey
> Rectangular Research
> http://www.rectangular.com/
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
> For additional commands, e-mail: java-dev-help@lucene.apache.org
>


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


Mime
View raw message