Invalid RGBA argument in matplotlib(python)

i just read csv and i want to plot3d by scatter3d.

import matplotlib.pyplot as plt

fig = plt.figure()
ax = plt.axes(projection='3d')
#ax.scatter3D(f2, f3, f1,'r')
ax.scatter3D(features[1:22,3], features[1:22,4], features[1:22,5],'r',cmap='viridis')

but..

ValueError: Invalid RGBA argument: masked_array(data=[0.12156862745098039, 0.4666666666666667,
                   0.7058823529411765, 0.378827099272238],
             mask=False,
       fill_value='?',
            dtype=object)

comes out... it looks like the data was,

data=[0.12156862745098039, 0.4666666666666667, 0.7058823529411765, 0.378827099272238]

which is made by pandas...

why the error comes out? I can not understand... the data is float and 1048575 rows × 1 columns in each features[:,3] or features[:,4]. May be the size is too big?? I tried [1:22,3] but nothing changed...