merging databases & multiple wordlists

[David Relson]

Hi all,

There have been a couple of questions about merging databases.  As y'all
know, "bogoutil -d wordlist.db > wordlist.txt" writes out all the
tokens, their ham and spam counts, and their timestamps.   When
"bogoutil -l wordlist.db < wordlist.txt" is run, it's really _adding_
the incoming numbers to the wordlist.  If wordlist.db doesn't exist,
bogoutil creates it and loads the new tokens into it.  If wordlist.db
_does_ exist, bogoutil adds new tokens and increases counts of tokens
that are already in the wordlist.  So, merging two wordlists can be done
with:

   bogoutil -d first.db > first.txt
   bogoutil -d second.db > second.txt
   cat first.txt second.txt | bogotuil -l new.db 

Bogofilter also supports message scoring using multiple wordlists.  The
effect of the above bogoutil commands can be realised by using the following:

   wordlist r,first,/dir/first.db,1
   wordlist r,second,/dir/second.db,1

Given the same precedence number, i.e. "1", for both wordlists,
bogofilter will check both wordlists.   If you also had:

   wordlist r,other,/dir/third.db,2

bogofilter would first look for the token in the "1" lists.  If found,
it wouldn't search any further.  If not found, it would try the "2"
list. 

HTH,

David