Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
Home -> Community -> Usenet -> c.d.o.server -> SQL Performance Optimization
Hi.
I have to SELECTs, both returning a key column in their result set. What I want is a join/intersection of both. The first result set is rather large, nonunique and comes from an expensive query so i want to avoid any sorting. The second ist just a SELECT ID FROM TABLE where the TABLE has an unique index.
I want to get the first rows ASAP.
The best strategy to get what i want is probaby to run along the first result set and to do lookups into the second table to see if the id is in it and to set a "have seen it flag". If I look up the same entry a second time I behave as if it is not there. This way i would get uniqueness without sorting.
This looks like a very unSELECTish strategy. Is there any way to write a SELECT that uses it (on 8.0.5).
Thanx,
Joachim
-- SELECT COUNT(1) FROM HIMMELSZELT.STERNE; Althoug this message has a valid From header, replies to user_at_netcologne.de where user = nc-zobeljo are preferred.Received on Sat Oct 14 2000 - 01:16:17 CDT