Re: get_time - negative time intervals ???

On Windows NT it seems that DBMS_UTILITY.GET_TIME is based on the value of the Win32 function GetTickCount (returns number of milliseconds since the OS was started), which is recycled each about 50 days. DBMS_UTILITY.GET_TIME is not guaranteed to keep on returning increasing numbers...

HTH. Finn

mminfo <mminfo_at_email.msn.com> wrote in message news:OIWmEBh$#GA.280_at_cpmsnbbsa02...
> We have a procedure that tracks the execution time of SPs by taking the
> difference between the SPs start and end times, as reported by get_time
> (dbms_utility).
>
> The problem we have been having is that this procedure is occaisionally
> netting us negative time intervals - i.e. time(END of SP) - time(START of
> SP) < 0!
>
> The problem is (seemingly) intermittent - it has happened more than once,
> however. There seems to be no correlation with time of day (incl. noon
and
> midnight), nor with any given SP, and we are certain that at least some of
> the SPs take more than 1/100 th of a second to execute, so the probability
> of some kind of rounding error seems small.
>
> Any explanations or tips would be very much appreciated.
>
> Thanks for your time,
> Mark
>
>
Received on Mon Sep 13 1999 - 18:02:03 CDT