Matplotlib: combine legend with same color and name

Question

If I repeat the plot with same color and label name, the label would appear multiple times:

from mpl_toolkits.mplot3d import Axes3D
fig = plt.figure()
ax = fig.gca(projection='3d')

x_labels = [10,20,30]
x = [1,2,3,4]
y = [3,1,5,1]

for label in x_labels:
    x_3d = label*np.ones_like(x)
    ax.plot(x_3d, x, y, color='black', label='GMM')

ax.legend()

Is it possible to make them into one, combing the same lables legends into one? Something like

I can produce the above pic by

from mpl_toolkits.mplot3d import Axes3D
fig = plt.figure()
ax = fig.gca(projection='3d')

x_labels = [10,20,30]
x = [1,2,3,4]
y = [3,1,5,1]
legend = False

for label in x_labels:
    x_3d = label*np.ones_like(x)
    ax.plot(x_3d, x, y, color='black', label='GMM')
    if legend == False:
        ax.legend()
        legend = True

But this feels very ugly, is there any good sotion? Or do I simply make the plot in a wrong way?

Answer 1

You should only show the label for one of the three sets of data. The can be done by adding an if/else statement in the label = ... in ax.plot(). Below is an example:

from mpl_toolkits.mplot3d import Axes3D
fig = plt.figure()
ax = fig.gca(projection='3d')

x_labels = [10,20,30]
x = [1,2,3,4]
y = [3,1,5,1]

for label in x_labels:
    x_3d = label*np.ones_like(x)
    ax.plot(x_3d, x, y, color='black', label='GMM' if label == x_labels[0] else '') 
    # above only shows the label for the first plot

ax.legend()

plt.show()

This gives the following graph:

EDIT:

If you have different colors then you could use the following to show the legend only once for each color:

fig = plt.figure()
ax = fig.gca(projection='3d')

x_labels = [10,20,30,40,50]
x = [1,2,3,4]
y = [3,1,5,1]

colors = ['black','red','black','orange','orange']
labels = ['GMM','Other 1','GMM','Other 2','Other 2']
some_list= []

for i in range(len(x_labels)):
    x_3d = x_labels[i]*np.ones_like(x)
    ax.plot(x_3d, x, y, color=colors[i], label=labels[i] if colors[i] not in some_list else '')

    if colors.count(colors[i])>1:
        some_list.append(colors[i])

ax.legend()

plt.show()

This gives the following graph: