Re: [Matplotlib-users] problem loglog plots with errorbars?

[John H. <jdh...@ac...>]

>>>>> "Istvan" == Istvan Szapudi <szapudi@IfA.Hawaii.Edu> writes:

    Istvan> Hi All, I am fairly new to matplotlib, and I am impressed
    Istvan> with its capabilities.

    Istvan> I have trouble plotting loglog plots with errorbars. I
    Istvan> have the following, program, a slightly modified version
    Istvan> of one given earlier by JDH on 9-28-2004 in this mailing
    Istvan> list:

This is definitely a gotcha that matplotlib needs to be smarted about
handling.  The log zero error is occurring when the transformation is
made on the axes limits and not on the data.  The autoscaler picks the
best min/max for the data coordinates, and will round down or up to
facilitate nice integer ticking.  When you plot with linear coords,
the autoscaler makes it's pick according to linear scaling, and then
when you change scales the old scaling is in effect and the log
transform fails when converting the viewport.

Solution: rescale the axes after changing coords, 
either manually

  ax.set_yscale("log")
  ax.set_xscale("log")
  axis([0.5*min(x), 2*max(x), 0.5*min(y-err), 2*max(y+err)])

or use the autoscaler

  ax.set_yscale("log")
  ax.set_xscale("log")
  ax.autoscale_view()

or set your log coords *before* calling plot

  ax = gca()
  ax.set_yscale("log")
  ax.set_xscale("log")
  errorbar(x,y,err,fmt='o')
  show()

and then the errorbar command will pick a "locator" to handle ticking
and viewport scaling appropriately from the outset.  This is the
approach taken in http://matplotlib.sf.net/examples/log_bar.py .

Until I get this fixed to work automagically, I'll make it a FAQ.

JDH