dropna parameter not working with pairplot?

[fizix00]

I have a dataframe called returns that has NaN values in the first row.

When I do sns.pairplot(returns[1:]) or sns.pairplot(data=returns.dropna()), there's no problem. However, when I try sns.pairplot(data=returns, dropna=True), I get the following error: ValueError: max must be larger than min in range parameter.

As far as I can tell, all three approaches should've yielded the same result.

[mwaskom]

No, unfortunately it doesn't. pairplot doesn't currently work well with dataframes that have heterogeneous missing data. There was some work on it (#409) but it stalled out.