Re: [SQL] Yet Another (Simple) Case of Index not used

Josh Berkus <josh@agliodbs.com> writes:
> In PostgreSQL:
> a) Put table_count_A on superfast media like a RAM card so that random seeks
> after 10,000 updates do not become a significant delay;

As long as we're talking ugly, here ;-)

You could use a sequence to hold the aggregate counter.  A sequence
isn't transactional and so does not accumulate dead tuples.  "setval()"
and "select last_value" should have constant-time performance.

            regards, tom lane