[rrd-users] Re: --width affects results?

[ralf-buero at kruedewagen.de]

On Tuesday 04 October 2005 13:40, Alex van den Bogaerdt wrote:
> On Tue, Oct 04, 2005 at 10:20:11AM +0200, ralf-buero at kruedewagen.de 
wrote:
> > > I'm sure a thourough analysis would explain the differences.
> > > For research, see footer.
> >
> > Yes, this is what I have already assumed during my studies. Even
> > if you use GPRINT, the results are also different depending on
> > image size. The results differ for me much more than in Grishas
> > example (for instance 98.1 vs. 96.0).
>
> It is not the image size itself that causes different data.
>
> If you have one-hour data available, and ask for 00:00 to 24:00,
> you get 24 rows of one hour each.
>
> If you have one-hour data available, and ask for now-24 to now,
> you get 26 rows of one hour each unless "now" happens to be on
> the hour exactly.  This is because you ask a different question.

Okay. Maybe this issue also affects the calculations. But if I define 
absolute start and end times, I get different results for different 
--width values.

> > The big problem is: How can I calculate a mathematically correct
> > average now based on all measurements between start and end time
> > ? Nobody (if not knowing the details of RRDtool) will trust me,
> > if I say, that averages (numbers) depend on the size of the
> > image. The size of the image should IMHO not affect numbers
> > calculated and printed by GPRINT or PRINT.
>
> You do not have a correct result if you do not start with correct
> assumptions.  From what I read here, you don't.

My assumptions seem to be correct for me, since I use absolute start 
and end times.

Ralf

--
Unsubscribe mailto:rrd-users-request at list.ee.ethz.ch?subject=unsubscribe
Help        mailto:rrd-users-request at list.ee.ethz.ch?subject=help
Archive     http://lists.ee.ethz.ch/rrd-users
WebAdmin    http://lists.ee.ethz.ch/lsg2.cgi