error bars with dataframe and pandas

[Hucky]

Hello everyone,

hope someone can give me an advice.

I have 3 dataframes with different values, each with time (x-values) and y values.
I calculated the average and the standard deviation for the error bars with

mean = dataframe.groupby(['time','y values']).agg({'y values':'mean'}).reset_index()
std = dataframe.groupby(['time','y values']).agg({'y values':'std'}).reset_index()
mean['err'] = std['y values']

Then I tried to plot everything with

fig, (ax1) = plt.subplots(1, 1) 
mean.plot(x='time', y='y values', style='.', color='b', legend=True, label='test', ax=ax1) 
ax1.errorbar(mean.index, mean['time'], yerr=mean['err'])
ax1.legend(loc='lower right', fancybox=True, fontsize=9)
ax1.yaxis.set_label_text("y values")  
ax1.xaxis.set_label_text("Time (h)")
ax1.set_title("test")  

But the error bars are not shown in the right way. I tried it already with mean.plot(...yerr=' '), but the same result. Someone have an idea?

[jefsummers]

What are you wanting, and what are you getting?
Check Matplotlib documentation

[Hucky]

Hi,

here you find the pictures
error bars should be and actual

the first one is, how the error bars should be and the second and third how they are.
I have absolutely no idea anymore.

[jefsummers]

Hard to tell, what is the shape of mean? std?
Have you tried making x and y in the errorbar call just mean and std, and no other arguments? Or mean.values and std.values and no other arguments?

[Hucky]

Found the mistake. I did
[pyhton]fig, (ax1) = plt.subplots(1, 1) ....[/[pyhton]]

and then later again
[pyhton]fig, (ax1) = plt.subplots(1, 1) ....[/[pyhton]]

Now I changes the second subplots ax1 to ax2 and everything works fine.
Thank you for your help.