secondary_y not always working

[jrkerns]

See a SO question I asked that has the relevant information: http://stackoverflow.com/questions/28840036/pandas-secondary-y-axis-for-boxplots

Essentially, the secondary_y argument of plot() doesn't do anything for boxplots and several other plot types. It will work for line plots however.

[TomAugspurger]

Should be helped by #9321 which will clarify what is / isn't supported.

I'm not sure that secondary_y makes sense for a boxplot though, which typically doesn't have a legend to indicate which elements are on the second axis. What do you think?

I'd recommend transforming your data before plotting.

[jrkerns]

I suppose if the secondary axis is tied to having a legend then the current implementation makes sense, and more documentation/warning may help. I agree if there is no legend then there is no explicit way the person viewing the plot knows what is plotted against what, but I guess I wish it was hackable so I could do it anyway and then explain it in the figure description.

Unfortunately for me, I can't transform the data as it's one of many plots I'll have for a publication and they all need to be the same style.

Thanks for the help.

[TomAugspurger]

Yeah, sorry to not be of more help. The only other thing I can think of is to append a (right axis) to the x-axis label, but this doesn't feel great.

There are certainly ways around this if you drop down to matplotlib though. I may have a chance later to take a crack at your SO questoin.