Hi,
I have a need for minimum memory footprint of the
index during search (would like to have it in RAM).
Good thing in the story, similarity calculation is not
necessary, only pure boolean model is OK.
I am sure I have seen somewhere one explanation from
Doug about disabling norms... but cannot find it now.
Could I implement my Similarity where all, idf, tf,
coord, lengthnorm return constant value... WIll this
disable loading of norms/positions/freqs in memory
during search? Also, if I use RAMDirectory in this
case (loading from FSDirectory) will these elements be
loaded in RAM?
Better expressed: If one needs only boolean model (no
score calculation needed), what shouls I do to
minimize memory usage and optimize speed.
My collection has a lot of very short documents (a
couple tokens per field), 5-7 fields, (approx 20Mio).
Indexing is not critical. Memory in runtime is just
about to fit in RAM.
thanks in advance, eksdev
___________________________________________________________
How much free photo storage do you get? Store your holiday
snaps for FREE with Yahoo! Photos http://uk.photos.yahoo.com
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
