Mailing List Archive: "Unique Entries" wrt sa-learn

On Sun, Mar 07, 2004 at 11:26:37AM -0700, LuKreme wrote:
> synced Bayes databases from journal in 0 seconds: 344 unique entries
> (354 total entries)

There were 354 journal entries. Of those, some were token atime updates.
Of those, there were tokens which occured more than once. The one with
the highest atime value is the only one that counts. So it looks like
you had 10 "duplicate" atime update entries, leaving 344 unique ones.

> synced Bayes databases from journal in 5 seconds: 7926 unique entries
> (7926 total entries)
>
> how could learning from 43 messages make the bayes journal jump from
> 344 unique entries to 7926?

Well, the rebuild took you from 344 to 0. Learning added more entries.
When you did the next rebuild, you had 7926 entries, none of which
were duplicate.

--
Randomly Generated Tagline:
"If you want to travel around the world and be invited to speak at a lot
of different places, just write a Unix operating system."
(By Linus Torvalds)