Re: Hom to force optimizer to use certaim index when doing update?

On Tue, 04 Nov 1997 22:13:09 GMT, tkyte_at_us.oracle.com (Thomas Kyte) wrote:

>On Tue, 04 Nov 1997 23:15:20 GMT, dtang_at_minn.net wrote:
>
>>Hi:

[snip]

>
>btw: the first query probably would have failed in real execution. Without the
>predicate on TableA, every row in A would have to have a match in TableB else
>the subquery would return zero rows but a subquery used with equals must return
>at least one and at most one row. The update would fail on the first row in A
>that didn't have a match in B.
>

Sorry, the above statement is wrong. The first query would have *worked* but the results probably would not be what you expected. If the subquery returned ZERO rows, field_1 in A would be set to NULL (which is probably NOT what you would have wanted).

The other part of the answer (use the predicate) was correct though.  

Thomas Kyte
tkyte_at_us.oracle.com
Oracle Government
Bethesda MD  

http://govt.us.oracle.com/ -- downloadable utilities  

[Quoted] NOTICE: In order to cut down on the amount of time I spend applying silly [Quoted] logic to email addresses like "In order to cut down on spam, remove THIS from my email" or "In order to cut down on spam, reverse my email address", etc, I will simply hit delete when that mail bounces back to me, which is what I do with spam myself.  

I haven't figured out whats more time consuming for me, deleting spam or resending emails after re-reading the whole thing to figure out the algorithm to apply. Received on Wed Nov 05 1997 - 00:00:00 CET