Zystkmt

I'm writing a script to draw plot via matplotlib, the following code is an analogy to my original script that reproduce the issue that I've encountered.

def func(ax, data, color, position):    # A function for plotting

    ax.plot(data[0], data[1], color=color)

    ax.spines[position].set_color(color)



fig = plt.figure()

data_1 = [np.linspace(0, 1, 10), np.linspace(0, 10, 10)]

data_2 = [np.linspace(0, 1, 10), np.linspace(10, 0, 10)]



ax = fig.add_subplot(111)

func(ax, data_1, 'r', 'left')



ax_1 = ax.twinx()

func(ax_1, data_2, 'b', 'right')



plt.show()

The expected plot should have both y axis colored, however, only right spine is colored, as shown below.

When zooming into left spine you can find a red shadow around y-axis, that means my colored spine is covered by another one, how to solve this issue by only modifying func()?

twinx does not only "twin" the y axis spine but all other three spines, too. So the red spine on the left is basically overdrawn (like you recognized already in your own answer). Instead of setting their color to None, you can set_visible() their visibility to False, which seems to be the preferred way, compared to lines 18-22 here.

So:

def func(ax, data, color, position):    # A function for plotting

    ax.plot(data[0], data[1], color=color)

    ax.spines[position].set_color(color)

    for pos in ['left', 'right']:

        if pos != position:

            ax.spines[pos].set_visible(False)

Seems like other spines would be created after executing this code: ax_1 = ax.twinx(), so I find a "dumb" solution:

def func(ax, data, color, position):    # A function for plotting

    ax.plot(data[0], data[1], color=color)

    ax.spines[position].set_color(color)



    if position == 'left':

        other = 'right'

    elif position == 'right':

        other = 'left'

    ax.spines[other].set_color('None')

Result:

This can solve my problem, but I'm still open for other beautiful solutions.