Re: Quick (?) query re Statspack parse:execute ratio calculation (Thick DBA !)

This is a little deceiving but the calculation is correct. They are using "1 -" in the formula ONLY to show in the report as 100% being the target value and 0% being the worst... hence 2.30% in your case.

If you take "1 -" out of the formula then 100% will be worst and 0% will be target value... as in your first calculation, 97.70% is now close to worst.

Regards
/Rauf

Norman Dunbar <Norman_at_DELETETHIS.BountifulSolutions.co.uk> wrote in message news:<cklls3$56u$1$8300dec7_at_news.demon.co.uk>...
> I'm having a brain block at the moment, and I need a gentle reminder of
> basic arithmetic :o)
>
> I'm looking at a 20 minute Statspack report and I *know* I've got
> parsing problems (no binds etc) but, I'm puzzled by the 'Execute to
> Parse %' calculation.
>
> Executions = 108,738
> Parses total = 106,240
>
> Calculation is 100 * (1 - parses/executions) which gives approx 2.30%.
>
> I suppose the question is 'what the hell does that really mean?'.
>
> If I want the percentage of parses to executions, I just use 100 *
> parse/executions and get 97.70% which immediately tells me that the
> application is crap because I want the figure to be a lot lower (I want
> 1 parse and 108,738 executions for example giving me 0.00092%) - I want
> parsing to be a tiny percentage of executions.
>
> If I want the percentage of executions to parses, I use 100 *
> executions/parses and get 102.35% which tells me that the application is
> still crap because I want executions to be very large percentage
> (108,738/1 * 100 = 10,873,800%).
>
> So, why all the subtractions from 1 ?
>
> (Please be gentle !)
>
>
> Cheers,
> Norm.
Received on Thu Oct 14 2004 - 12:33:00 CDT