Python – Vectorization conversion of decimal integer arrays to binary arrays in numpy

I’m trying to convert an array of integers to their binary representation in Python. I know native python has a function called bin that does this. Numpy has a similar function: numpy.binary_repr.

The problem is that these are not vectorization methods because they only take one value at a time. So in order to convert the entire input array, I have to use a for loop and call these functions multiple times, which is not very efficient.

Is there any way to perform this conversion without using a for loop? Are there any vectorized forms of these features? I tried numpy.apply_along_axis without success. I’ve also tried using np.fromiter and map, but it doesn’t work either.

I know similar questions have been asked several times (e.g. here). ), but none of the answers given are actually vectorized.

Pointing me in any direction would be appreciated!

Thanks =)

The easiest way is to use binary_repr and vectorize, which will preserve the original array shapes, for example:

binary_repr_v = np.vectorize(np.binary_repr)
x = np.arange(-9, 21).reshape(3, 2, 5)
print(x)
print()
print(binary_repr_v(x, 8))

Output:

[[[-9 -8 -7 -6 -5]
  [-4 -3 -2 -1  0]]

[[ 1  2  3  4  5]
  [ 6  7  8  9 10]]

[[11 12 13 14 15]
  [16 17 18 19 20]]]

[[['11110111' '11111000' '11111001' '11111010' '11111011']
  ['11111100' '11111101' '11111110' '11111111' '00000000']]

[['00000001' '00000010' '00000011' '00000100' '00000101']
  ['00000110' '00000111' '00001000' '00001001' '00001010']]

[['00001011' '00001100' '00001101' '00001110' '00001111']
  ['00010000' '00010001' '00010010' '00010011' '00010100']]]