Re: Can relational alegbra perform bulk operations?

From: Walter Mitty <wamitty_at_verizon.net>
Date: Thu, 01 Oct 2009 10:23:29 GMT
Message-ID: <Bq%wm.3890$tl3.2127_at_nwrddc01.gnilink.net>


"Clifford Heath" <no.spam_at_please.net> wrote in message news:jcTwm.13128$Bl2.1664_at_newsfe14.iad...
> Walter Mitty wrote:
>> When code jockeys set out to learn SQL, the hardest thing to get them to
>> learn is to refrain from coding in such a way as to gove implicit "hints"
>> to the optimizer in the way they express their queries.
>
> Unfortunately MySQL is still teaching the new generation that the order of
> joins actually matters.
>

I didn't know that. Sigh.

>> Getting programmers to think in terms of "what" rather than "how" has
>> been a continuing struggle since the 1950s.
>
> Now you're singing my song. This is one of the most fundamental problems
> of the industry, going far beyond just the database world.
>
> I've had to use hints in MSSQL (2000) in order to avoid a nasty optimiser
> bug where it ignores the obvious index that can make a join work. This was
> with a DELETE FROM a WHERE NOT EXISTS(...) (an anti-semi-join).
>
> --

I once ran across an optimizer that didn't realize that "group by A, B" and "group by B, A" are equivalent, and thereby failed to use a composite index.

That optimizer was otherwise very, very good. I seldom worried about speed. Received on Thu Oct 01 2009 - 12:23:29 CEST

Original text of this message