[FIXED] Matplotlib Line Overlap/Resolution

Issue

I'm making a plot in Matplotlib where it is imperative for different lines to not overlap when they have differing x-values. For some reason, when I draw lines, there is overlap if they are close enough together (but not overlapping). For example,

fig = plt.figure(num=None, facecolor='w', edgecolor='k')
ax = fig.add_subplot(2, 1, 1)
ax.plot((0, 10000000), (3, 3), linewidth = 2, markersize = 0, clip_on = True, aa = True)
ax.plot((10000001, 200000001), (3, 3), linewidth = 1, markersize = 0, clip_on = True, aa = True)

plt.savefig('test.png', format='png')

I want the blue line and the green line to not overlap at all, since 10000000 < 10000001.

I made the blue line slightly thicker, so if you zoom in, you can see how the blue and green lines overlap. I made the dpi very high, so it isn't a resolution problem. I will have to be dealing with large numbers because I am working with genome data--are the large numbers an issue? When I plot with x-values of (0 to 10000) and (10001, 20000), the problem still exists.

Thanks so much for your help.

Solution

It has to do with the cap style of the Line2D object, the default style is 'projecting', which result in overlapping, see a zoom-in-ed PDF:

We want to change it to 'butt' style:

L1=ax.plot((0, 10000000), (3, 3), linewidth = 2, markersize = 0, clip_on = True, aa = True)
L2=ax.plot((10000001, 200000001), (3, 3), linewidth = 1, markersize = 0, clip_on = True, aa = True)
for item in L1+L2:
    item.set_solid_capstyle('butt')

The gap is very small, sure, since it is 1/10000000.

Or if you want, a quick dirty solution is just to draw a small white circle marker of size=1 at (10000000.5, 3).

---

• matplotlib.pyplot.plot can use kwargs from matplotlib.lines.Line2D, which includes the parameter solid_capstyle
• Here is an additional example setting solid_capstyle in the plot call

fig, ax = plt.subplots(1, 1, figsize=(8, 4))

L1 = ax.plot((0, 1), (1, 1), linewidth=10, label='No capstyle')
L2 = ax.plot((1, 2), (1, 1), linewidth=7, label='No capstyle')


L3 = ax.plot((0, 1), (0.9, 0.9), linewidth=10, solid_capstyle='butt', label='With butt capstyle')
L4 = ax.plot((1, 2), (0.9, 0.9), linewidth=7, solid_capstyle='butt', label='With butt capstyle')

ax.set_ylim(0, 2)
ax.grid()
ax.legend()

Answered By - CT Zhu