2

Okay this is werid, i keep getting the error, randomly.

ValueError: matrix must be 2-dimensional

So i tracked it down, and cornered it to basically something like this:

a_list = [[(1,100) for _ in range(32)] for _ in range(32)]
numpy.matrix(a_list)

Whats wrong with this? If i print a_list it is clearly a 2d matrix of tuples, however numpy does not believe so.

UberJumper
  • 18,745
  • 17
  • 66
  • 86

3 Answers3

4

The easiest way around this is to just use a numpy array, instead of a numpy matrix:

a_list = [[(1,100) for _ in range(32)] for _ in range(32)]
arr=numpy.array(a_list)

Numpy matrices are strictly 2-dimensional, and a_list is 3-dimensional. So numpy matrices are not an option.

Community
  • 1
  • 1
unutbu
  • 777,569
  • 165
  • 1,697
  • 1,613
1

tuples have more than one value, so they are considered a dimension. So you're creating a 3d matrix.

nosklo
  • 205,639
  • 55
  • 286
  • 290
0

What are you trying to do? In numpy, a matrix is a 2-d array of numbers.

You can make a 32x32x2 "matrix" with numpy.array(a_list). You could also make a 32x32 array of tuples using numpy object arrays, but there is little that would be better for than the 32x32x2 case (since you can think of it as a 32x32 array of 2-element arrays).

Andrew Jaffe
  • 25,476
  • 4
  • 47
  • 58