I was brought here for a reason not explicitly mentioned in the answers so far, so to save others the trouble:
The error also occurs if the function arguments have changed order - for the same reason as in the accepted answer: the positional arguments clash with the keyword arguments.
In my case it was because the argument order of the Pandas set_axis
function changed between 0.20 and 0.22:
0.20: DataFrame.set_axis(axis, labels)
0.22: DataFrame.set_axis(labels, axis=0, inplace=None)
Using the commonly found examples for set_axis results in this confusing error, since when you call:
df.set_axis(['a', 'b', 'c'], axis=1)
prior to 0.22, ['a', 'b', 'c']
is assigned to axis because it's the first argument, and then the positional argument provides "multiple values".