[thelist] Doing things just once -- tracking in database.

[Bill Moseley]

I have an "account" table with a date column "end_date" that indicates
when the account expires.

I need to send out email to accounts that are about to expire.  I want
to be able to track which accounts have been sent the email.

One simple solution is to have an additional boolean column on the
account table "reminder_email_sent", but that does not seem very
scalable as more and more events will need to be tracked.

I'm leaning toward an event table with columns "account_id",
"event_id", and "processed_time" so that for any event I can find
accounts that may be ready for the event (e.g. find all accounts
that will expire in less than 10 days and do not have a row in the
event table for the event within, say, the last 30 days).


Does this seem like a sane approach?  Obviously, this is a common
task for applications so I'm wondering if there are better solutions.

Do you use a table to track processing events?  What columns do you
include in that table?


Thanks for your input,


-- 
Bill Moseley
moseley at hank.org
Sent from my iMutt