> Do not use 3.0 with PostgreSQL. It's known to have problems. Actually, > the problem you're seeing might be DBD::Pg related. Did you possibly > just install a new version? Make sure you are using the version listed > in sql/README.bayes.
I haven't noticed anything out of the ordinary after I made the switch to SQL bayes, that is, after I created a few indexes to help. DBD::Pg is version 1.21. I haven't messed around with Perl settings or packages lately to cause this. More than likely I've just now hit upon the bug described in this earlier version of DBD::Pg in sql/README.bayes for some reason. But what is it about just lately that's causing bayes to fail? Is it a record in a table that somehow was inserted that's causing all subsequent lookups to fail? > There is custom PostgreSQL support in 3.1, which just had RC1 released, > you can try that one. It works very well and is 7-27x faster than the > code in 3.0. I read about the custom PostgreSQL support in the new SA and am very excited to hear about it. I've been following some of the discussion about the problems with Postgres support in 3.0 early on. I'd would much rather wait until the production release of 3.1 -- and can live without bayes for a while -- than running a RC in production. Andre Nicholson
