Update percent_return.py enforce fill method on pandas pct_change

[bossdown123]

adjust the pct change

[twopirllc]

@bossdown123

Curious.

Does the following not calculate percent change correctly with the exception of filling some nan ?

pr = (np_close / roll(np_close, length)) - 1
pr[:length] = nan
# Why not handle nans here?
pct_return = Series(pr, index=close.index)
# Or here?

Thanks,
KJ

[twopirllc]

@bossdown123

Upon further review, there does not appear to be any difference between either calculation or speed on the development branch. Thus I am keeping the original code unless you have further evidence to the contrary.

Here is the core code tests I ran in Jupyter:

• pd.testing.assert_series_equal(tapr, pdpr, check_names=False) never threw False either.

Thanks for helping to making the library better!

Kind Regards,
KJ