Well, this might be easier:
1) Generate a png image for each time series you want to look at.
For example, lets do one for today, and one for ysterday.
2) Now take all of the data lines from the first graph, and superimpose
them onto the second.
Repeat if you want more than two graphs as the composite.
This works, and you can see an example perl script here:
http://newover.consumption.net/broncplus/composite.html
Essentially, the program uses PerlMagick to walk the first image. For
each pixel, check to see if we should copy it to the next image. It's
not that fast, doing a 650x350 image in 5seconds, but seems reasonable.
torleiv
On Fri, 2 Nov 2001, David Gabler wrote:
>
> Clifton,
>
> Thank you for confirming my thoughts. Strange I never though of it as
> different time offsets (the obvious way to explain it). That probably would
> have helped my archive searches.
>
> David
>
> -----Original Message-----
> From: Clifton Royston [mailto:[EMAIL PROTECTED]
> Sent: Wednesday, October 31, 2001 3:05 PM
> To: David Gabler
> Cc: [email protected]
>
> > I figure I can do this if I do dump, pull the data from the dump I want
> to
> > graph, turn to proper xml format, restore the data to a rrd and then
> graph
> > 4 diff DS's.
> >
> > Just curious if there is an easier way.
>
> That sounds like it would work, and no, there probably isn't an easier
> way.
> -- Clifton
>
>
--
Unsubscribe mailto:[EMAIL PROTECTED]
Help mailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdmin http://www.ee.ethz.ch/~slist/lsg2.cgi
