Re: How are joins performed when using a database link?

Rob Cowell wrote:

> DA Morgan wrote:
> >
> > Rob Cowell wrote:
> >
> > > <snipped>
> > >
> > >
> > > > it often happens here that guys spend more time typing out a request
> > > > for advice than actually having a go, making a few mistakes & testing
> > > > things out.
> > >
> > > I think you are being a bit harsh on this poster.
> >
> > You are correct with respect to how it works. But incorrect with respect to the
> > basic premise which is the impact on performance.
> >
> > In less time than it took to post the request and wait for a response the OP
> > could have run a test. And while not knowing "how it worked" would sure know
> > the relative performance on THEIR system.
> >
> > Daniel Morgan
>
> So he would have known what improved performance on this particular
> query on this system but not why, or how to find out why.
> He would have been back to square one with the next problem of this
> type, and in no position to think about how to solve a more complex
> distributed problem.
>
> I appreciate there are a lot of people who don't seem to be able to find
> a freely available manual but I don't think this topic is particularly
> well explained anywhere obvious.
>
> Let's face it most problems here can be solved if you want to plow
> through Metalink and make some empirical discoveries.

Lets try it your way. Lets assume an engineer from Oracle, one that wrote the underlying C code sent him a copy of the source code along with full documentation on how it works.

Of what possible value would that be? The only things that matter are performance, scalability, and stability. I don't know the underlying engineering differences between an Intel P2 CPU and an Intel P4 CPU. But I can sure tell you which one I want to use to support an application.

Lets let this one drop but while I agree that it is nice to know ... how ... it has limited utility.

Daniel Morgan Received on Mon Mar 17 2003 - 13:57:03 CST