Hi,
I've exercised db according to the pattern described in the
previous mail. The surprising result (I still have to double check)
is that your method (a node for each word occurence) eat less space
than storing location data for unique words. And it's 3 times faster
than updating or doing put duplicate with updated data + delete old one.
I'll send more precise results when I'm finished. I created a bench
directory with a program to exercise various aspects of db related
to our purpose. I'll not commit that ;-)
One of the things I tried is to check the efficiency of
prefix compression. But I can't make it work. db_stat constantly
says that no prefix saving occurs despite the fact that my 500 000
keys are all prefixed with the OOOOOOOOOO string. I've been running
in circles during the past hour. I'm starting to wonder if the code
works :-( Looking at the db code is rather difficult.
I'd be very gratefull if you can just send me the output
of a db_stat on a file of yours that shows that prefix compression
indeed work. I'd be even more gratefull if you said 'ha, ha, there
is a trick, you must be carefull to do the following ....'. BTW,
I'm using Database.h interface to create and manipulate the db
file when runing the benches.
Cheers,
P.S. Sorry in advance if this mail is out of sync with previous mail
you may have sent, I've not had the opportunity to check my mail today.
--
Loic Dachary
ECILA
100 av. du Gal Leclerc
93500 Pantin - France
Tel: 33 1 56 96 09 80, Fax: 33 1 56 96 09 61
e-mail: [EMAIL PROTECTED] URL: http://www.senga.org/
------------------------------------
To unsubscribe from the htdig3-dev mailing list, send a message to
[EMAIL PROTECTED] containing the single word "unsubscribe" in
the SUBJECT of the message.
