Question

I have some 3D data e.g. d=[x, y, z, f]
where z is a column of numbers in Z, used as color information.

f is a flag which is

1. 0 if x and y have some specific values (ugly^^)
2. 1 if x and y are ok

So for the good data d[ d[:,3] == 1 ] I want to generate a profile

plt.imshow(resampled.T, extent=extent, vmin=MIN, vmax=MAX, origin='lower')

and for the ugly data d[ d[:,3] == 0 ] I want to just use a specific color, e.g. black

Is there a way to realize that?

EDIT: Combining the comments of @eumiro and @Rutger Kassies, I have now the following result

Which is satisfying I think.
For the sake of completeness (or maybe there are some optimization I’m not aware of^^), here is the code and the data:

import numpy as np
from matplotlib.mlab import griddata
import matplotlib
import matplotlib.pyplot as plt


def plotprofile(x, y, z0, name='dummy', save=1):
    #plt.figure()
    N = 50j
    z = z0[:,0]
    extent = (min(x), max(x), min(y), max(y))
    xs,ys = np.mgrid[extent[0]:extent[1]:N, extent[2]:extent[3]:N]
    resampled = griddata(x, y, z, xs, ys)

    cmap = plt.get_cmap()
    cmap.set_bad(color = 'k', alpha = 1.)
    #plt.imshow(resampled.T, cmap='Greys', extent=extent, origin='lower', interpolation='spline36')
    plt.imshow(resampled.T,  cmap=cmap, extent=extent, origin='lower',  vmin=min(z),  vmax=-min(z),interpolation='spline36')

cbar=plt.colorbar()
s=20
plt.ylabel(r"$y$", size=s)
plt.xlabel(r"$x", size=s)
plt.xlim([x.min(),x.max()])
plt.ylim([y.min(),y.max()])

if save:
    for end in ["pdf", "png", "eps"]:
        print "save %s.%s"%(name,end)
        plt.savefig("%s.%s"%(name,end))
else:
    plt.show()
plt.clf()


if __name__ == '__main__':
    filename = 'data.txt'
    data = np.loadtxt(filename)
    x = data[:,0]
    y = data[:,1]
    z = data[:,3:]
    plotprofile(x, y, z,  'dummy', 0)

Answer 1

Can’t you create just a normal colourmap in z by using f to mask z?

dd = d[:, :3]
dd[:,2] = dd[:,2] * d[:,3]

Then convert to an image like this:

M = dd.max(0)
m = dd.min(0)

x = np.arange(m[0], M[0] + 1)
y = np.arange(m[1], M[1] + 1)
[X, Y] = np.meshgrid(x, y)

Z = np.zeros_like(X)

for num in range(0,size(dd, 0)):
    Z[dd[num, 0], dd[num, 1]] = dd[num, 2]

Now you should be able to plot Z like a normal image or as a surface against [X, Y]