Andrew Sullivan <andrew@libertyrms.info> writes:
> But you _could_ write a trigger which inserts into a "staging" table,
> and write a little daemon which only updates the count table with the
> data from the staging table. It's a mighty ugly hack, but it ought
> to work.
The $64 question with this sort of thing is "how accurate (up-to-date)
does the count have to be?".
Given that Josh is willing to vacuum every five minutes, he might find
that returning pg_class.reltuples is Close Enough (TM).
regards, tom lane