Re: why CTAS faster than Insert select?

From: Dave <x_at_x.com>
Date: Tue, 29 Mar 2005 18:56:49 GMT
Message-ID: <Rxh2e.7948$Ab.7224@text.news.blueyonder.co.uk>

"linda" <linglipeng_at_yahoo.com> wrote in message news:1112121974.648540.12120_at_l41g2000cwc.googlegroups.com...
> Hi,
>
> I've been testing the performance between Create Table As Select (CTAS)
> vs Insert Select for large volume of data (40 million records or so).
> Both in nologging mode, parallel, etc. It consistently shows that CTAS
> faster than Insert Select. My understanding is that, if you parallel
> both create/insert process and select process, then theoratically they
> should perform about the same (same amount of work, same number of
> parallel processes, minimum redo and undo...). Could anybody explain
> why this is the case?
>
> Thanks,
> Linda
>

lets see you queries and the stats then Received on Tue Mar 29 2005 - 12:56:49 CST