Word list size question
David Relson
relson at osagesoftware.com
Sun Oct 31 20:22:30 CET 2010
Hello Doug,
The size of the wordlist database varies with the amount of training
you do. 120MB doesn't seem particularly large. I use Berkeley DB,
rather than SQLite and my wordlist is 3x larger than that.
Dumping and reloading a wordlist will compact it -- at least for a
while. With Berkely DB as tokens are added, the DB will (as
needed) split a database block into 2 blocks to provide room for the
new token.
Dump/reload is done with bogoutil and bogofilter's source distribution
includes a script a bf_compact for this purpose. Look to see if
your distro includes the script. As I don't use sqlite, I can't
predict how much space will be recovered by running the script.
HTH,
David
On Sat, 30 Oct 2010 22:28:12 -0700 (PDT)
Doug wrote:
> I have been running bogofilter for about 2 weeks and my wordlist size
> is now almost 120M
>
> 117240 -rw-rw---- 1 admin admin 119927808 Oct 31 01:04
> wordlist.db
>
> at what rate can I expect this to grow? Does it just continually
> grow? If so I suppose at some point it becomes unusable?
>
> I am using Sqlite. Is there a way to compact it using Sqlite?
More information about the Bogofilter
mailing list