Python: multithreading with concurrent.future missing data compare to multiprocessing

Question

Problem:

I have a list which contains 500 url inside, in each url it could contains one or more code (my desired output dataframe is one code