Asymmetrical errorbar with pandas [duplicate]

Question

I want to plot asymmetrical errorbars with pandas. According to official docs this should work

df = pd.DataFrame([[1,0.2,0.7]]) fig, ax = plt.subplots() df[0].plot.bar(yerr=[df[1], df[2]], ax=ax)

But pandas renders errorbar as df[1] for both lower and upper limits (-0.2/+0.2 istead of -0.2/+0.7):

Where do I make a mistake?

I use pandas v0.20.3 with python v2.7.13 under Windows 7.

John Zwinck · Accepted Answer · 2017-08-21 13:44:50Z

4

Your yerr is 1D:

yerr=[df[1], df[2]]

It needs to be 2D, specifically one row per data point and each row having two values for negative and positive error:

yerr=[[df[1], df[2]]]

answered Aug 21, 2017 at 13:44

252k44 gold badges346 silver badges459 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

actually either 1D or 2D seems to output the correct result in a newer version of pandas, I tested with pandas 1.2.4 and matplotlib 3.3.4