Generating an unknown number of lists for a python function

Question

Continuing the conversation in Can a variable number of arguments be passed to a function? , I'd like to learn how to generate a random number of arguments for a function in python 2.7.

In particular, while looping over various categories, I would like to pass an unkown number of array_like (or lists?) to scipy.stats.f_oneway.

A simple example that works would be:

list_male = [34.316349, 34.32932, 34.27, 34.33905, 34.328951]
list_female = [34.61984, 34.34275, 34.6389, 34.44709, 34.51833]
f_oneway(list_male, list_female)

This works yielding

F_onewayResult(statistic=12.15815414713931, pvalue=0.0082363437299719927)

because I knew that my category gender has only two classes male, female.

But what if I am running a loop of many categories and I do not want to predetermine specific lists? E.g, if a category column animal in a pd.DataFrame has an unknown number of classes. I would like within the loop to do something like df['animal'].unique().tolist() and then create array_like (or whatever required) to feed to f_oneway.

Would `f_oneway(*a_list_of_lists)` work? You can try with `a_list_of_lists = [list_male, list_female]`. — 9000, Feb 23 '17 at 16:08

score 4 · Accepted Answer · edited May 23 '17 at 12:16

I'm unfamiliar with f_oneway(...), but assuming that it can take a variable number of arguments, and that you have a variable number of lists, you can do this:

list_male = [34.316349, 34.32932, 34.27, 34.33905, 34.328951]
list_female = [34.61984, 34.34275, 34.6389, 34.44709, 34.51833]
list_both = [...]
list_neither = [...]
list_of_lists = [list_male, list_female, list_both, list_neither]

f_oneway(*list_of_lists)

This is accomplished with the use of the splat operator, which unpacks the list contents into individual arguments

Generating an unknown number of lists for a python function

1 Answers1