Re: New SqlPlus type utility available for download

Christopher.Jones_at_oracle.comX wrote:

> Daniel Morgan <dmorgan_at_exesolutions.com> writes:
> > Now that you bring it up there has been a long standing
> > problem with SQL*Plus ... in fact so long standing that
> > I've gotten tired of mentioning it. What is wrong with
> > providing useful timing information when someone does SET
> > TIMING ON?
>
> By 'useful' do you mean including statement parsing timing?
> Or is there an issue with the documentation or output
> messages?
>
> > I've yet to meet anyone that can accurately interpret the
> > results (well perhaps excepting Tom Kyte and we haven't
> > met).
>
> Solaris (and Windows since about 8.0 or 8.1) give a single
> elapsed time from end to end:
>
> SQL> set timing on
> SQL> execute dbms_lock.sleep(5);
>
> PL/SQL procedure successfully completed.
>
> Elapsed: 00:00:05.33
> SQL>
>
> The string is hours:minutes:seconds, with seconds being
> shown to two decimal places. The extra 0.33 of a second in
> this example is the overhead of SQL*Plus, the network and
> the server (I'd just started the database).
>
> The output is port specific, depending on the low level
> timing functions available. By now (meaning now that UNIX
> libraries and ports have settled down), most UNIXes should
> have the same output as Solaris. I don't know what message
> is printed on platforms I don't have much contact with e.g
> z/OS.
>
> Chris
>
> --
> Christopher Jones, Oracle Corporation, Australia.

[Quoted] How about just something that matches the format HH:MI:SS.DD

No one, outside of a few experts, has any idea what to do with something where any of the above except perhaps the decimals following seconds, is three digits in length. It is not uncommon to see 00:14:107 whatever that's supposed to mean.

Daniel Morgan Received on Wed Aug 28 2002 - 18:17:01 CEST