Re: Can we optimize an ORDER BY clause?

Have you looked at Oracle clusters? If you have only one table per cluster, you can approximate the behavior that you mention. An added advantage is that if you have two or more tables in the cluster your joins on that index key perform very fast.

Richard

Vincent Qing Yin <vyin_at_cs.ubc.ca> wrote in article <5aua3d$ok_at_harpoon.cs.ubc.ca>...
> Here's a small question for Oracle programmers.
>
> I come from a Sybase back ground. I noticed that Oracle doesn't have
> the concept of a "clustered" index. In Sybase, a clustered index
> forces the physical order of records on disk to be the same as the
 logical
> order of the index. In a SELECT statement, if I use an ORDER BY clause
> on the columns that have the clustered index, then the server immediately
> start to return rows without actually performing the sort (because the
> sort would be unnecessary.)
>
> In Oracle, is there a way I can optimize an ORDER BY clause so that
> the server will not build an intermediate result set for sorting?
>
> Thanks for any advice.
>
> --
>
> Q Vincent Yin | Repeat
> vyin_at_cs.ubc.ca | delete(next->bug);
> (604) 876-9096 (H) | Until 0 = 1;
>
  Received on Wed Jan 08 1997 - 00:00:00 CST