Python: How to flatten list of lists where each sublist is a contains a single float

Question

I want to flatten my list that looks as follows:

generated = [[108.88114502], [19.29647502], [4.08611068], [52.33578872], [134.54672018], [14.22388561], [22.04375228]]

To something like this:

flat_list = [108.88114502, 19.29647502, 4.08611068, 52.33578872, 134.54672018, 14.22388561, 22.04375228]

I tried the following but I get a TypeError saying floats are unsubscriptable:

flat_list =  [item for sublist in generated for item in sublist]

Can someone help me out please?

Answer 1

After OP edits and comments

• TypeError saying floats are unsubscriptable when using [item for sublist in generated for item in sublist] means the list elements are single values and single values in a list (e.g. [[14.22388561], [22.04375228], 34.6576])
• Check if each element is a list.
  • If item type is a list, take the value at index 0, assuming each list contains only 1 value.
  • If item type is not a list, take item

generated = [item[0] if type(item) == list else item for item in generated]

Solution from before OP edited the question

• I'm guessing generated is a np.array, so you can use .reshape
  • If a np.array is printed, the output looks like that shown by the op, a list with no ,
  • The array shown, has the shape (7, 1), and you want it reshaped to (1, 7)

g = generated.reshape(1, 7)

# display g
array([[108.88114502,  19.29647502,   4.08611068,  52.33578872,
        134.54672018,  14.22388561,  22.04375228]])

• If generated is a list, converting it to a np.array and using reshape will work.

generated = np.array(generated)

generated = generated.reshape(1, 7)  # where 7 is the length of the list)

%%timeit comparison

Given g as a list of lists

import numpy as np

np.random.seed(365)
g = [[np.random.randint(100000)/100] for _ in range(1000000)]

Test

%%timeit
[item for sublist in g for item in sublist]
74.6 ms ± 415 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)

%%timeit
np.array(g).reshape(1, 1000000)
243 ms ± 2.51 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)

Given g as a (1000000, 1) np.array

np.random.seed(365)
g = [[np.random.randint(100000)/100] for _ in range(1000000)]
g = np.array(g)

Test

%%timeit
[item for sublist in g for item in sublist]
867 ms ± 3.11 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)

%%timeit
g.reshape(1, 1000000)
289 ns ± 1.16 ns per loop (mean ± std. dev. of 7 runs, 1000000 loops each)

Answer 2

flat = [l[0] for l in generated]