Bar chart: x-axis order incorrect for Int64Index with multiple columns

[mc51]

My conda env.
There seems to be an issue with the order of the x-axis in some situations. This happens when the DataFrame index is of type Int64Index and we depict more than one value in a bar chart.

This works as expected, i.e. the order on the x-axis is 1,2,10,11:

data = pd.DataFrame([[5, 1],
                     [5, 2],
                     [6, 10],
                     [7, 11]], columns=['a', 'b']).set_index('b')
data.hvplot.bar()

However, when adding a column to the DataFrame there is an issue. The order of the x-axis will be wrong, i.e. 1,10,11,2. So it seems that the axis gets interpreted as str instead of int:

data = pd.DataFrame([[5, 1, 66],
                     [5, 2, 66],
                     [5, 10, 66],
                     [5, 11, 66]], columns=['a', 'b', 'c']).set_index('b')
data.hvplot.bar()