Is reduce method requiring initializer?

Question

I am trying to append an arbitrary number of PySpark dataframes together. This is attempted with the union_all function below:

from functools imp