Re: Performance of count(*)

Tino Wildenhain wrote:
>> You guys can correct me if I'm wrong, but the key feature that's
>> missing from Postgres's flexible indexing is the ability to maintain
>> state across queries. Something like this:
>>
>> select a, b, my_index_state() from foo where ...
>> offset 100 limit 10 using my_index(prev_my_index_state);
>>
>
> Yes, you are wrong :-) The technique is called "CURSOR"
> if you maintain persistent connection per session
> (e.g. stand allone application or clever pooling webapplication)

That's my whole point: If relational databases had a simple mechanism for storing their internal state in an external application, the need for cursors, connection pools, and all those other tricks would be eliminated.

As I said earlier, relational technology was invented in an earlier era, and hasn't caught up with the reality of modern web apps.

> If its a naive web application you just store your session
> in tables where you can easily maintain the scroll state
> as well.

One thing I've learned in 25 years of software development is that people who use my software have problems I never imagined. I've been the one who was naive when I said similar things about my customers, and was later embarrassed to learn that their problems were more complex than I ever imagined.

Craig