Re: Unexpected Tkprof Row Count

From: Matt <mccmx_at_hotmail.com>
Date: 2 Sep 2004 01:11:29 -0700
Message-ID: <cfee5bcf.0409020011.65255937@posting.google.com>

> We can see I can make it 100 as the result of the sort/group by step
> (or any number I feel like)...
>
> but the important thing to note is the cr= numbers (which you won't
> see in your version). they show the addition cr's in this second
> query were 100% due to the larger "TABLE ACCESS FULL OBJ#(63160)" --
> it wasn't that we ran the view 100 times for the nested loops outer --
> but that seems to be what contributed to the wrong statistics count.

OK, so the huge row count for the SORT(GROUP BY) is a red herring.. Thanks very much for your explanation.

I will need to look elsewhere in the tkprof output to determine where my time is being spent. I was convinced that this huge SORT operation was responsible for most of my 25 second response time, but it seems that I was focusing on the wrong step.

Matt Received on Thu Sep 02 2004 - 03:11:29 CDT