Re: Eliminating cartesian merge

Terry Dykstra wrote:
> Couldn't help but notice:
> AND ( job.effdt >= TO_DATE (TO_CHAR (SYSDATE, 'YYYY-MM-DD'), 'YYYY-MM-DD')
>
> Why so complicated? I would code:
>
> AND ( job.effdt >= TRUNC(SYSDATE)
>

I've noticed that too, but this is delivered Peoplesoft code that we're not supposed to change.

If I just remove the distinct from the main query, it executes in seconds. There are only about 5000 rows returned by the non-distinct version of th equery. Why is simply adding a distinct increasing the the time from a few seconds to minutes to only sort 5000 rows?

-- 
To reply by email remove "_nospam"