How to change layers of display for the plt.errorbar function

When plotting data via plt.plot and plt.errorbar it happens that the plt.errorbar has a higher layer of display, meaning that it is always plotted in front of plt.plot() data.

For using:

plt.plot(x, data_1, color='r')

plt.errorbar(x, data_2 ,yerr= std_2, color='b')

The errorbars and value points overlap the red data as expected, because the errorbar was plotted afterwards:

plt.errorbar plotted after plt.plot

But if you now use

plt.errorbar(x, data_2 ,yerr= std_2, color='b')

plt.plot(x, data_1, color='r')

The points of the plt.errorbar() are still a layer of display before the plt.plot() data:

plt.plot plotted after plt.errorbar

especially when looking at the whole plot (zooming out) the plt.plot() data with already unreasonably big markersize can be barely recognised:

zoomed out/whole plot

I tried using plt.scatter instead of plt.plot, but the layer of display problem is the same here

for both (plt.plot and plt.scatter) I clearly expected, that their layer of display is higher when putting them below the plt.errorbar() function in the code

You should use the barsabove argument of the errorbar. If you just do, e.g.:

import numpy as np
from matplotlib import pyplot as plt


x = np.linspace(0, 4, 15)
y = x
yerr = 0.5

fig, ax = plt.subplots()
ax.errorbar(x, y, yerr=yerr, mfc="b", marker="o", ms=15, ls="none", capsize=2) 
ax.plot([1.5, 2.5], [1.5, 2.5], marker="o", ls="none", mfc="r", mec="r", ms=50)

you will get something similar to what you are seeing:

If you add barsabove=True, you will instead get:

x = np.linspace(0, 4, 15)
y = x
yerr = 0.5

fig, ax = plt.subplots()
ax.errorbar(x, y, yerr=yerr, mfc="b", marker="o", ms=15, ls="none", capsize=2, barsabove=True)
ax.plot([1.5, 2.5], [1.5, 2.5], marker="o", ls="none", mfc="r", mec="r", ms=50)

You can also investigate using zorder for fine grained control of the plotting order.