Re: Can relational alegbra perform bulk operations?

From: Walter Mitty <wamitty_at_verizon.net>
Date: Wed, 30 Sep 2009 20:20:15 GMT
Message-ID: <34Pwm.3811$tl3.1_at_nwrddc01.gnilink.net>


"Banana" <Banana_at_Republic.com> wrote in message news:4AC3B9AE.4040100_at_Republic.com...
> 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.
>
> You're preaching to the choir here. I do cringe a little when I see all
> those fancy "hints" which I've never really seen the occasion to use. I'm
> sure it may make sense in a exceptional or specialized case but that's all
> it is. Most queries are best well left enough to the engine.
>
>> Getting programmers to think in terms of "what" rather than "how" has
>> been a continuing struggle since the 1950s.
>
> Fascinating; it shows that more things change, the more they stay same.
>
>> Matters were made even worse when programmers learned how to manipulate
>> the "rules based optimizer" in Oracle, before the days when it had a cost
>> based optimizer. The "cool" programmers learned how to tell Oracle how
>> to do its job by doing such things as putting a certain table last in the
>> FROM clause. All of this distracted from coming up with a good logical
>> design and clear coherent code. And "hints" were an even deeper descent
>> into "how" rather than "what".
>
> Well, as I agreed with you, one shouldn't really be fiddling with the
> engine, but as I remarked in an earlier post if we were to port a complex
> database project from say, SQL Server to Oracle, and even if we didn't use
> any vendor-specific features, several SQL rewriting may be warranted
> simply because of performance differences due to the differences of how a
> SQL statement is parsed & optimized between two engines.
>

I'm not going to take exception to all attempts to optimize. I do take exception to a premature preoccupation with performance optimization, to the detriment of all other measures of "goodness". I didn't mean to preach to the chior, here. I just wanted to point out what this issue looks like from the trenches. Received on Wed Sep 30 2009 - 22:20:15 CEST

Original text of this message