lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Morus Walter <>
Subject Re: warm up lucene, especially sort by cache
Date Wed, 02 Mar 2005 09:50:55 GMT
Chris Lu writes:
> 1. Need an efficient way to pick up the most frequent words in an index.
>     I think this can be done, any example will be appreciated.
I don't see an alternative to looping through all terms and look at their

> 2. search by the most freqent words, with sort by options
> Is this the only way to warm up lucene? For large indexes, the first 
> sort-by search is slow.
that's independent of what you search for.
sort by is done by creating a in memory array of all field values of all
documents. This array is cached and reused for further searches.

So if you use sort, doing one sort after creating the index might be useful.

For reading relevant parts of the index into OS caches, I'd rather use
the most commonly searched terms, than the most frequent ones.


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

View raw message