Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Supporting multiple oracle versions in a trigger

Re: Supporting multiple oracle versions in a trigger

From: Jack Addington <jaddington_at_shaw.ca>
Date: Fri, 21 Oct 2005 18:04:39 GMT
Message-ID: <X4a6f.248962$oW2.60777@pd7tw1no>

> WHen you get the book, read the chapter on locks and latches.  Then you
> will see exactly where the scalability problem will occur.  Until then I
> will not be able to explain it properly.

Ok well spent the morning googling and reading, and will come back to this after my book arrives, but I can't drop this yet for a few days. I'm going to assume for now that all the scalability issues are with the 'client tables' that I am dropping/creating on the fly and not with my raw data storage.

Given my constraints:

Is this a better solution given the worse case scenario?:

Before dataCollection change:

ViewClientPrivateT = select data in transpose format. TableClientPrivateT = indexed table essentially CTAS from viewT viewClientPublicT = select * from TableT DataCollection col links data collection to TableClientPrivateT

So when the underlying collection changes to require a rebuild:

Then have a cleanup process at 2am to drop any objects no longer linked to data collections (TableClientPrivateT/ViewClientPrivateT);

Apart from the workload on the database to populate the new table, I see the main concurrancy issues as:
- A select is being processed from viewClientPublicT while I attempt to
create/replace it. (would a drop/create of a synonym to the viewClientPublicT be better?)

Just some thoughts...

thx

JAck Received on Fri Oct 21 2005 - 13:04:39 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US