how to change format of my error bars in matplotlib module

Question

I am trying to figure out how to make only upper error bar with cap "-" in my plot, which I already figured out by lolims argument. Unfortunately, my error bar is marked with an arrow, and I would prefer the default one. The upper subplot is with error bar I would like to have, and the lower one is with the mark I would like to have on my upper subplot.

My code:

import pandas as pd
import matplotlib.pyplot as plt

sheet = pd.read_excel(load file with my data)

fig, axs = plt.subplots(2, 1)
fig.suptitle('Gene expression', fontsize=16)

axs[0].bar(sheet.columns, sheet.iloc[17],hatch="/", color="white", edgecolor="black")
axs[0].errorbar(sheet.columns, sheet.iloc[17], yerr=sheet.iloc[22], capsize=3, ecolor='black', fmt=' ', elinewidth=1,
                lolims=True)

axs[1].bar(sheet.columns, sheet.iloc[17],hatch="-", color="white", edgecolor="black")
axs[1].errorbar(sheet.columns, sheet.iloc[17], yerr=sheet.iloc[22], capsize=3, ecolor='black', fmt=' ', elinewidth=1,)
plt.show() 

and picture of my plots:

How can I achieve it?

Answer 1

You have three possibilites depending on whether you'd like to see the lower caps:

1. Specify an (2,N)-shaped yerr or
2. Plot the error bars behind the bars
3. Change the caps afterwards (as commented below by BigBen)

import matplotlib.pyplot as plt

fig,ax = plt.subplots(3)
x = range(1,4)

ax[0].bar(x, x, fc='w', ec='k')
ax[0].errorbar(x, x, yerr=[[0,0,0], [.1,.2,.3]], fmt='none', capsize=3)

ax[1].bar(x, x, fc='w', ec='k', zorder=1 )
ax[1].errorbar(x, x, yerr=[.1,.2,.3], fmt='none', capsize=3, zorder=0)

ax[2].bar(x, x, fc='w', ec='k')
_, cl, _ = ax[2].errorbar(x, x, yerr=[.1,.2,.3], lolims=True, fmt='none', capsize=3)
cl[0].set_marker('_')
#cl[1].set_marker('')  # to remove the lower cap