Purely by accident, we found out that a SELECT from DUAL was much faster.
So we replaced all tables in our application's SQL by DUAL.  Guess what?

It all works so fast now we can hardly see the screen results!

Yet I still ran an EXPLAIN PLAN on our SQL.  We are still doing the dreadful
TABLE SCANS.

Does anybody know of a good way to index the DUMMY column in DUAL
so we can stop this FULL TABLE SCAN nonsense?
Any ideas on sizing, tablespace split and so on for even better performance?

TIA

Yes, we tried the same approach with similar results.  If you reformat your
hard drives you should have no more problems.

You know, most users underestimate the importance of mass-storage
lubrication. Try a little drive-spindle grease before you go through all the
trouble of re- formatting - you'll probably notice 50 - 100% improvement,
depending on the age of your drives.

>Does anybody know of a good way to index the DUMMY column in DUAL
>so we can stop this FULL TABLE SCAN nonsense?
>Any ideas on sizing, tablespace split and so on for even better performance?
>
We've always just added '... and rownum = 1' to all the SELECT's; that way
only the first row is read, rather than the whole table.

G.