[ILUG] SQL question

[Donncha O Caoimh]

Niall O Broin wrote:
> 
> Can anyone who's running MySQL 3.23.xx verify if COUNT(DISTINCT column)
> works as advertised and as per standard SQL. If that's the case, it's
> finally time to upgrade from 3.22.32 :-) (I was waiting for subselects, but
> I think Godot will come first)

Tried that query on a 3.23.32 server and it returned a figure. Leaving
out the COUNT() bit it returned the same number of records.. :)

> 
> Talking of waiting for things, I have often considered changing to Postgres
> because of its greater compliance to standard SQL, but I've been put off
> becuase MySQL has the reputation of being much faster. Lately, however, the
> Postgres people having been putting out benchmarks which show Postgres as
> being faster than MySQL (well, they would, wouldn't they). Now I know that

I want to try that too, and might get around to it in the next 2 months.
(I'd be lucky!)
What interests me is that updates/inserts can happen while large selects
go on, which I believe PostgreSQL can do to a certain extent. The
PHPBuilder article on it tested CVS PostgreSQL with MySQL 3.23(alpha)
and found PostreSQL to be much faster.. 
If anyone is testing, be sure to read up on tuning PostgreSQL, I've read
that in the default install it sucks for performance (does syncing every
few seconds, etc..)

Oh yeah, Niall, you should have rung one of us, there was plenty of beer
on Saturday, and plenty of quotes for the quotes page which probably
shouldn't be repeated in public ;)

Donncha.