Re: Tkprof73 question

In article <961092758.12840.0.pluto.d4ee154e_at_news.demon.nl>,   "Sybrand Bakker" <postbus_at_sybrandb.demon.nl> wrote:
>
>
> [snip biggest part of the explanation]
> > The 'disk' statistics show that 100 disk blocks were read from the
> > datafiles for the session. The fact that c query buffers were used
> > probably indicates that the same data blocks were continually
 revisited
> > throughout the session, not the best situation.
> >
> > David Fitzjarrell
> > Oracle Certified DBA
> >
> >
> > Sent via Deja.com http://www.deja.com/
> > Before you buy.
>
> May I ask you a question?
> Are you sure about this. I always learned to interpret this as
> 12329 buffers used *of which only 100* where physical reads.
> You probably agree that's quite a different way of reading it, in my
 opinion
> this would still be an inefficient statement given the ratio of buffer
 gets
> and records, but not with respect to physical I/O
>
> Regards,
>
> Sybrand Bakker, Oracle DBA
>
>

I went back to my Oracle documentation, to see if I was, indeed, mistaken, and I am not. The disk column of the tkprof output table indicates the number of disk blocks read and the query column reports the number of buffers retrieved. Simply because data blocks are revisited does not indicate an I/O problem, and I never stated such. I did indicate that this was not the most efficient scenario. It was not my intent to infer the problem was I/O related.

David Fitzjarrell
Oracle Certified DBA

Sent via Deja.com http://www.deja.com/
Before you buy. Received on Thu Jun 15 2000 - 00:00:00 CDT