Re: cannot have subqueries in CONNECT BY clause

Alexander V. Silantiev doodled thusly:

>One more question.
>If view like
>create or replace view classify_rview as
>select cid,cmaster from classify f
>where
>exists(select null from classifyrw where cid=f.cid and
> username=user)
>and exists(select null from classifyrw2 .........)
>and exists(select null from classifyrw3 ........)
>........................
>and exists(select null from classifyrwN ........)
>;
>
>Whould be the difference in throughput
>between the above method and the next joins
>
>select from classify,classify2,,classify3....,,classifyN
>

Hehehe! I knew there was a catch!

:-)

Actually, if they all have indexed keys, there should be no diff. The optimizer should be smart enough to only select data from the first table. Because you only actually select columns from that one. Then again, it's been known to go South every now and then. Particularly if there are many classifyrwn tables...

Cheers
Nuno Souto
nsouto_at_optushome.com.au.nospam Received on Mon Nov 19 2001 - 07:19:45 CST