Brandon Craig Rhodes <brandon(at)oit(dot)gatech(dot)edu> writes:
> The `working' test case omits the
>     AND (account, policy) NOT IN
>      (SELECT account, policy FROM policy_accounts_active)
> condition from the end of executor_active, which magically makes the
> executor_hamlet rule start firing as it should.

I don't think this is a bug.  The executor_hamlet rule fires after the
executor_active rule does; therefore the (account, policy) pair *has
already been inserted into policy_accounts_active*, and will be found
when executor_hamlet re-executes the select to look for it.

My advice to you is to use triggers, not rules, for pushing data from
one table to another; especially when you need logic this complex to
decide what to do.

			regards, tom lane

In response to

• bug? rules fail to cascade after NOT IN at 2003-03-03 19:43:30 from Brandon Craig Rhodes

Responses

• Re: bug? rules fail to cascade after NOT IN at 2003-03-04 03:23:43 from Brandon Craig Rhodes

pgsql-hackers by date

Next:	From: Brandon Craig Rhodes	Date: 2003-03-04 03:23:43
	Subject: Re: bug? rules fail to cascade after NOT IN
Previous:	From: Christopher Kings-Lynne	Date: 2003-03-04 01:49:59
	Subject: Re: problem importing languages in CVS tip