Re: Overall executions at DB level

From: Ana C. Dent <anacedent_at_hotmail.com>
Date: Thu, 10 Aug 2006 13:45:39 GMT
Message-ID: <Xns981B44C68A0DDanacedenthotmailcom@70.169.32.36>

"Gints Plivna" <gints.plivna_at_gmail.com> wrote in news:1155217209.219563.143570_at_m73g2000cwd.googlegroups.com:

> Ana C. Dent wrote:

>> "Spendius" <spendius_at_muchomail.com> wrote in
>> news:1155213159.808984.286920 @b28g2000cwb.googlegroups.com:
>>
>> > Hi,
>> > Do you think doing
>> > SQL> SELECT SUM(executions) FROM V$SQL[AREA]
>> > at regular intervals relevantly reflects the number of total
>> > SQL executions whatsoever in your database ?
>> > (I'm not sure as I've noticed sometimes that substracting
>> > a value from the next one returns a negative number
>> > - guess it's because in the meantime Oracle rid V$SQL
>> > of a few statements...)
>> >
>> > Thanks.
>> > Spendius
>> >
>> >
>>
>> SO?
>> Not all SQL is created equal.
>> SELECT SYSDATE FROM DUAL;
>> is much less resource intensive than a SELECT doing a 12 table join.
>> By your methodology they both would count the same.

>
> And you offer what?
> Giving some ratio to each SQL statement? :)
>
> Gints Plivna
> http://www.gplivna.eu
>
>

Yes, BCHR would be as useful a measure as SUM(executions).

And you offer what? Received on Thu Aug 10 2006 - 08:45:39 CDT