[ILUG] Running spamassassin on a large number of email messages

[Justin Mason]

Timothy Murphy writes:
> On Tuesday 09 September 2008, FRLinux wrote:
> 
> I have a directory containing about 50,000 email messages,
> of which I am sure over 95% is spam.
> 
> What is the best (quickest) way of passing these messages through spamassassin
> and deleting those that fail the test.
> [I'm running spamd on the machine in question.]
> 
> A simple script - if appropriate - would be very much appreciated.
> 
> > Hello Timothy, you didn't specify the mailbox format (mbox or Maildir)
> 
> They are in maildir format - 50,000 separate files.


You could run 

    spamassassin -t /path/to/maildir/new > mbox1
    spamassassin -t /path/to/maildir/cur > mbox2

It'll output mboxes with the filtered versions of each message.  You then
need to replace the maildir with the mboxes and remove the ones that
match "X-Spam-Flag: YES".

--j.