> From: Alex Murzaku [mailto:[EMAIL PROTECTED]]
> 
> Is there an easy way to index a list of words with 
> precomputed frequencies
> (WORD->FREQ)? Or should I create an artificial file 
> containing each term
> WORD repeated FREQ times? I guess I should modify DocumentWriter and
> manipulate the postings...

For indexes containing more than a few thousand documents, tokenization is
not the dominant cost of adding documents.  So creating artificial documents
is actually not that inefficient of a way to go, and it doesn't require much
code.

Doug

_______________________________________________
Lucene-dev mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/lucene-dev

Reply via email to