Hi there,

I am trying to plot a surface using plot_surface.

As this surface is generated using griddata, the corresponding z-values

is a masked array.

When I plot this masked array using plot_surface,

the colormap is completely upset, the whole surface appears in blue.

One can do something like

plot_surface(x,y,z.filled(z.mean()),cmap=cmap.jet)

but this is not really what one would like to have, because the missing

values are being

displayed, and in this case, the missing values are certainly not equal to

the mean of the surface.

Alternatively, it would be nice if griddata allowed to extrapolate a bit out

of the convex hull

of the given data. It seems that natgrid is able to do this (I found options

that specifies this), I installed natgrid, but in the matplotlib interface

to natgrid, one cannot specify any options proper to natgrid,

and extrapolation seems to be switched off, leading to masked arrays being

returned.

One more alternative would be that the color calculation for the surface

would only be done

for the non-masked fields in the masked array, but I do not know how to tell

matplotlib to do this

or how to do this by hand.

Any help appreciated!

## ยทยทยท

--

View this message in context: http://old.nabble.com/plot_surface-masked-array-tp27266471p27266471.html

Sent from the matplotlib - users mailing list archive at Nabble.com.