Re: [Matplotlib-users] Bug in colorbar()

[Eric F. <ef...@ha...>]

On 2012/07/26 7:52 PM, Jeffrey Spencer wrote:
> I am trying to make a plot with a colorbar that has a reduced axis over
> which the colorbar is executed.
>
> This is set via passing in a norm to contourf:
>          logNorm = colors.Normalize(vmax=0,vmin=-100)
>          surf = ax.contourf(X,Y,logZ, map_scale, cmap=cm.jet, norm=logNorm)
>
> The output of this will have the colorbar extend to the full range of
> the data and not limited by the norm set:
>
> cbar = fig.colorbar(surf, shrink=0.70, aspect=36, fraction=.15,pad=.03)
>
> so I assumed modifying by setting the norm like this would do the trick:
>
> cbar = fig.colorbar(surf, shrink=0.70, aspect=36, fraction=.15,pad=.03,
> norm=logNorm)
>
> This isn't what happens. norm has no effect. The norm is recognized but
> not passed to ColorbarBase is my guess from doing this instead to get
> the desired effect:
>
>          axcb, _ = mpl.colorbar.make_axes_gridspec(ax, shrink=0.7)
>          cbar = mpl.colorbar.ColorbarBase(axcb, cmap=cm.jet, norm=logNorm)
>
>
>
> Is this a bug or any reason why the norm is not passed through if
> specified in colorbar??

Yes, there is a reason why it is not passed through; both the cmap and 
the norm are taken from the mappable, so that the colorbar will actually 
match the mappable.  There may be a bug, but it is not immediately 
obvious to me.  Please provide a complete, minimal, self-contained 
example that illustrates the problem, and specify what version of mpl 
you are using.

Eric