Oracle FAQ Your Portal to the Oracle Knowledge Grid

Home -> Community -> Usenet -> comp.databases.theory -> Re: Problems with SELECT *

Re: Problems with SELECT *

From: Bob Badour <>
Date: Sat, 15 Mar 2003 00:22:03 -0500
Message-ID: <7ezca.76$>

"Pablo Sanchez" <pablo_at_dev.null> wrote in message news:Xns933E8086A9F5Fpingottpingottbah_at_216.166.71.233...
> "Bob Badour" <> wrote in
> news:HXpca.47$
> > Why, then, did you cut the prior context from the post?
> I believe that people should use their news readers and news as it was
> designed to be used: cut out the relevant pieces and if someone wants
> to walk up the parent thread, they can do so. Think of it as a form
> of removing redundant data.

News services have some peculiarities making such reliance inadvisable. Messages do not necessarily arrive in chronological order and do not necessarily remain on news servers for very long. Or necessarily arrive at all, for that matter.

When I cut out previous context and address a remark to a specific statement, I usually intend to communicate that the removed context is irrelevant and to focus my remark to the addressed statement.

Nevertheless, I have reviewed the thread a couple times now, and I don't see what provoked your objection to the perfectly valid observation you addressed: Choosing to keep the dbms ignorant of actual requirements and asking for more than needed prevents the dbms from using effective physical optimizations. The observation is as true for requesting too many columns as it is for requesting too many rows. I don't see how you jumped from that observation to the conclusion that this somehow causes applications to rely on anything.

I think it was a perfectly valid observation, and I still do not understand your objection to it.

> > And how does the context of the original poster suggest anything
> > about application reliance on indexes?
> Interesting, that's not what I intended to write and when I re-read
> what I wrote, I can see how it could have been misperceived:
> A gentle reminder, the original poster was stating the reasons why
> to avoid 'select *' and putting down the above reason is a good
> reason but I don't think it's paramount. My vote is to avoid it but
> primarily because people's times are more expensive than machine.
> I should have put a 'comma' after the 'select *' to force the break
> between the OP and the other person's point on piggybacking.

Even if the reason is not paramount, it is an important reason. I still do not understand your objection to it. How does requesting what is actually needed cause an application to rely on anything but its own requirements?

> > I don't see that anyone in the entire thread suggested anything
> > about reliance on indexes.
> Exactly and that's why I was talking about the intention of the
> original poster who has a list of reasons. I wouldn't put
> piggybacking as a key item in the list, more of an ancillary point.

Okay. So you would agree that it belongs on the list? Somewhere?

Would you agree with the general principle that telling the dbms the precise accurate requirements better allows it to do its job? Including physical performance optimization?

Can you explain how specifying an accurate list of required data causes an application to rely on anything other than its own requirements? Or have you now changed your mind about that too?

> Also, why are you so aggressive? Is this simply your writing style or are
> you just angry for some reason?

Aggressive? Angry? I am only trying to understand your objection. Your denials and dismissals do not help any. Are you, by any chance, perceiving other peoples' innocent contributions to this thread as some kind of personal challenge or affront? Received on Fri Mar 14 2003 - 23:22:03 CST

Original text of this message