> > What is it that I'm missing? > > > > bayes_store_module Mail::SpamAssassin::BayesStore::SQL > > If you truly missed it, no prob, but if it wasn't obvious from the > documenation please let me know so I can make it more obvious.
*sigh* Yeah, I really missed it. Maybe what I need to make it more clear to me is flashing neon pointers that say "Hey Ron, you missed this part!" WRT to the comments that PostgreSQL seems slow-- I'll certainly dig into it if I end up having the same experience. I probably won't have autolearn turned on though, so it may not be quite the same. If anybody else is able to duplicate the results or is having the same issue, the postgres devs usually end up asking all/some of the following when people report performance issues: 1) VERSION 2) Have you done a recent "vacuum analyze" (especially after many inserts) 3) If the query is still slow, what is the output of "explain <put-the-sql-query-here>" say about the plan that postgreSQL is following? My questions: 4) What is the key for the btree? If the 'id' part of the index (defined as " bayes_token_pkey primary key btree (id, token)" is a sequence, then there may be issues with getting an unbalanced btree as an index. I don't recall if that resulted in performance issues or index growth (or both), and whether or not it was fixed in more recent versions. Then of course there's the page that Bruce Momjian wrote about performance tuning for postgreSQL: http://www.postgresql.org/docs/aw_pgsql_book/hw_performance/ One thing that was confusing to me as I started to look in to the queries-- how are the tokens created? Looking at the tokens that 'make test' puts into the database don't seem to match any of the words (or things that I would think of as tokens) in the data/spam/001 test file. (I discovered this only because make test failed the bayessql tests 17 and 22(?). None of this is critical to me at the moment-- it does seem to recognize the ham and spam I feed it and score it appropriately when I scan it again so something is working right. Thanks for the great response! -ron
