Re: Oracle versus MS Sqlserver 2000: Technical Comparison of the Features in the two databases

"Howard J. Rogers" wrote:

> On Mon, 10 Mar 2003 08:27:50 +0000, DA Morgan wrote:
>
> > "Howard J. Rogers" wrote:
> >
> >>
> >> > 4. PERFORMANCE and TUNING
> >> >
> >> > a. No control of sorting (memory allocation) in Sql Server.
> >> > Oracle can fully control the sort area size and allows it
> >> > to be set by the Dba.
> >>
> >> Have I introduced you to PGA_AGGREGATE_TARGET yet?? It's the way of the
> >> future, so this particular point is moot.
> >
> > Unless you know something I don't I disagree with your conclusion.
> > PGA_AGGREGATE_TARGET allows automation: It doesn't force it. From what I have seen
> > Oracle is making is possible for DBA's to choose their
> > path from among two. One path is for a SQL Server type ease of management but the
> > other seems to be
> > provding the ability for hands-on control.
>
> You must have missed the phrase "the way of the *future*". IE, of course
> you have two ways of doing it *for now*. What's the betting that in
> version 13.6 sort_area_size will be a deprecated parameter, and in version
> 15.2 it will be obsoleted?
>
> No, I don't know anything more about this than you do. It's merely wild
> speculation. I'm also wildly speculating that automatic undo management
> will be compulsory around version 11.6, and ASSM will be compulsory about
> the same time.
>
> >> > f. A Sql-Server dba claimed that fully qualifying the name of an object
> >> > in Sql Server code would lead to performance gains of 7% to 10%.
> >> > There are no dictionary performance problems like that in Oracle.
> >>
> >> Clearly, you've never had a plethora of public synonyms. There *can be*
> >> performance problems related to such dictionary issues.
> >
> > Lets not turn an MS bashing into an Oracle bashing. As you said earlier bad DBAs
> > and
> > developers can do bad things. A plethora of public synonyms is just an example of
> > the obvious.
>
> I wasn't Oracle bashing (I wouldn't have a job for long if I did!). I was
> rebutting the claim that 'there are no dictionary performance problems' in
> Oracle. There *can* be. Yes, it takes dodgy design to do it. But that
> minor qualification wasn't being made by the original lose.... sorry,
> poster.
>
> >
> >> > 5. MISSING OBJECT TYPES IN SQL SERVER
> >> >
> >> > a. No public or private synonyms
> >> >
> >> > b. no independent sequences
> >>
> >> This really annoys me, frankly. Its apples and oranges time again. Why
> >> don't you save yourself some time and just post a message saying 'SQL
> >> Server is crap'? I actually like the SQL Server 'identity'
> >> auto-incrementing data type. I wish Oracle had something similar. The lack
> >> of sequences in SQL Server is not a 'missing object type' if you consider
> >> that they have something else which does the job very nicely, and rather
> >> more simply than the Oracle equivalent.
> >
> > Unless you are striving to create systems that serialize why would you prefer
> > autonumbering
> > to sequences? Can you reverse number with autonumering? Autonumber with only
> > multiples
> > of a seed number? Truncate a table, reload it, and restart numbering at any number
> > you choose?
> > Use a single numbering to sequence inserts into multiple tables?
> >
> > There is a lot more to sequences than just pulling a number of the top of a stack.
> >
>
> Absolutely. But I'd like the choice. When I *am* in 'pull number off the
> top of a stack' mode, the identity data type would do just fine. If I want
> things to go sideways and backwards, I'd chose sequences.
>
> Is there a serialization on identities? Yes. Could they not create a
> cached identity, just as we have cached sequences? Yes. If they chose.
>
> For now, so long as I aware of the performance issue, I'd like the
> opportunity to use an identity in Oracle if one were available.
>
> Regards
> HJR
Thanks for the clarification though I did assume you were not bucking for a spot in the unemployment line.

The main reason I will continue to lobby against autonumbering is that almost everyone will choose that initially because it is the easy out. Then when performance and scalability suffer they will either slow Oracle down to a SQL Server-like pace or they will have to redesign.

We suffer enough from the dumbing down of people in many walks of life. Sequences are not more than a few keystrokes:

CREATE SEQUENCE xyz;

all other parameters being unnecessary

AND sequence_name.NEXTVAL

But they at least don't dumb down Oracle developers. If Oracle gets as brain-dead as Access it will find Access is a competitor.

Daniel Morgan Received on Mon Mar 10 2003 - 16:53:02 CST