Manish Khettry wrote:
The qeury plan was erroneusly reporting an index being
used when none was being used. Try to do a distinct scan
on a table without an index.
---------
maximumdisplaywidth 20000;
create table noindex (x int);
call SYSCS_UTIL.SYSCS_SET_RUNTIMESTATISTICS(1);
select distinct x from noindex;
values SYSCS_UTIL.SYSCS_GET_RUNTIMESTATISTICS();
---------
Okay, thanks for the example--I see your point. It doesn't seem like we should
be printing an index here after all...
Let me know if this clarifies or makes you (or any
other commiter) feel better (or worse) about the patch?
I guess I feel better now :)
Thanks for the quick reply,
Army
