Convert mixed list of Numpy arrays and Python lists to only Python lists

Question

I have a list, d2_sum, that contains numpy arrays and python lists, and I'm trying to convert all to python lists so I can perform some slicing operations. I have,

d2_sum_list = [i.tolist() if type(i) == 'numpy.ndarray' else i for i in d2_sum]

This executes with no errors but does not convert any of the numpy arrays to python lists. What am I missing here?

you need to remove the quotes around numpy.ndarray, since it's a class and not a string — asdf101, Apr 14 '21 at 20:33
Or just `[list(x) for x in d2_sum]`. Sure, you needlessly re-create some lists, but it probably has a negligible impact on performance and keeps the code cleaner. — juanpa.arrivillaga, Apr 14 '21 at 20:36
@juanpa.arrivillaga: `list(x)` doesn't do the same thing as `x.tolist()`. `tolist()` on a 2D array will produce a list of lists, while `list(x)` will produce a list of arrays. Also, `tolist()` produces ordinary Python scalars rather than NumPy scalars. — user2357112, Apr 14 '21 at 20:44

asdf101 · Answer 1 · 2021-04-14T20:40:05.093

1

remove the quotes, because type(something) == "numpy.ndarray" will always be False:

d2_sum_list = [i.tolist() if type(i) is numpy.ndarray else i for i in d2_sum]

edited Apr 14 '21 at 20:40

answered Apr 14 '21 at 20:35

asdf101

569
5
18

More idiomatically, you'd use `type(i) is numpy.ndarray`, or `isinstance(i, numpy.ndarray)`, although those are not exactly the same (although effectively the same here) – juanpa.arrivillaga Apr 14 '21 at 20:37
very good point actually, I tend to forget `is` exists unless I check for None – asdf101 Apr 14 '21 at 20:39

Convert mixed list of Numpy arrays and Python lists to only Python lists

1 Answers1