Hi,

At November 22, 2004, 09:10, Adomas Urbanavicius wrote:

> The point is , that procedure itself does more reads, than is expected.
> For example proc P1 :
> After I called P1, apears dataset DP1 with count reads from tables
> for example 100.000, and result rows =10;
> If I wish to order it (select * from P1 order by xxx ), it should order
> dataset DP1, and reads should remain 100.000.


> But instead it creates dataset DP2 , with reads > 250.000 !
> This is the point, that stored procedure with order by clause causes big
> OVERHEAD, and there is no escape from that.

Which tool are you suing to get these statistics?
And, do you notice a difference in performance(time wise) between the
two(without looking at those numbers)?

If I'm not mistaken, those statistic numbers that you are stating are
not returned by the server, but are calculated by the tool you are
using.

--
Best regards,
Daniel Rail
Senior System Engineer
ACCRA Group Inc. (www.accra.ca)
ACCRA Med Software Inc. (www.filopto.com)