Hi Tony, What you describe is the correct and expected behaviour. When zooming on the x or/and the y axis, the quantities in the stats box must be recomputed from the bin contents. When displaying an unzoomed histogram, the stats box can use the statistics quantities computed at fill time, but this not possible (by definition) on a zoomed histogram. Rene Brun On Thu, 1 Aug 2002, Tony Colley wrote: > Fellow ROOTers: > > I have displayed a scattergram. The display is shown in the attached > before-canvas.gif. > > I then use the mouse/cursor to zoom in on a portion of the x axis. The > statistics within the displayed statistics block automatically generated > by ROOT change! See the new statistics (for the exact same set of > points) in the attached after-canvas.gif. The change may not be drastic, > but it is disconcerting. If one had instead used the mouse/cursor to > zoom in on a portion of the y axis, the exact same behaviour is noted... > that is, the statistics change and they change to the same new > statistics as when zooming on the x axis. When I unzoom, the statistics > go back to the original statistics. > > Now, if I zoom in to where not all of the points are within the zoomed > range, the statistics change again! Specifically, the number of entries > remains constant, but the remaining numbers change. If I continue to > zoom in to reduce the number of points within the displayed range, > sometimes the statistics change and sometimes they don't. > > I noticed this behaviour two or three months ago, and it still occurs > with my current ROOT. My current configuration is ROOT 3.03/06 (built 10 > July 2002) on a Linux RedHat 7.2 workstation (Dell OptiPlex GX110) using > gcc 3.0.2. > > Another issue: sometimes ROOT refuses to zoom to the range I've drawn > out with the mouse/cursor. No matter how many times I try, it simply > refuses to change. It is NOT anywhere near the resolution of float or > anything... for example, in this scattergram, once I've zoom the y axis > into the range 40 to 41.6 (which is the y range of the data points), it > does not want to zoom in any more... no matter how large or narrow a > band I try to zoom in on. There are many other unexpected behaviours > with axis zooming, but this is enough for one email. > > Cheers, > Tony Colley > >
This archive was generated by hypermail 2b29 : Sat Jan 04 2003 - 23:51:02 MET