New matrices

The goal

Making a new matrix…

Using import numpy as np is the standard.

Simple example – new np.zeros()

Define the size of your new matrix with a tuple, e.g.

M = numpy.zeros((DIM_0, DIM_1, DIM_2, …))​

1d

import numpy as np

M = np.zeros((2))
print(M)

Output:

[0. 0.]

2d

import numpy as np

M = np.zeros((2, 3))
print(M)

Output:

[[0. 0. 0.]
 [0. 0. 0.]]

3d

import numpy as np

M = np.zeros((2, 3, 4))
print(M)

Output:

[[[0. 0. 0. 0.]
  [0. 0. 0. 0.]
  [0. 0. 0. 0.]]

 [[0. 0. 0. 0.]
  [0. 0. 0. 0.]
  [0. 0. 0. 0.]]]

Simple example – recycle np.zeros_like()

If you have a matrix with the same size you want then you can use zeros_like. This will also copy other properties like the data type.

as a prototype use

N = numpy.zeros_like(M)

import numpy as np

M = np.zeros((2, 3, 4))

N = np.zeros_like(M)
print(N)

Output:

[[[0. 0. 0. 0.]
  [0. 0. 0. 0.]
  [0. 0. 0. 0.]]

 [[0. 0. 0. 0.]
  [0. 0. 0. 0.]
  [0. 0. 0. 0.]]]

Remember unpacking

This is an optional topic!

import numpy as np

d = (3, 4)
M = np.zeros((2, *d))

print(M)

np.empty is not np.zeros

If you are sure that you don’t care about what is inside the matrix in the beginning use

M = numpy.empty((DIM_0, DIM_1, DIM_2,...))​

Empty claims a region in the memory and uses it for a matrix. Zeros goes one step further. It fills the memory with zeros.

Thus random junk (i.e. data that was stored prior at that memory position) with be the content of a matrix if you use empty. However, np.empty() is faster than np.zeros().

​import numpy as np
M = np.empty((10, 4))
print(M)

[[1.66706425e-316 0.00000000e+000 6.89933729e-310 6.89933730e-310]
 [6.89933729e-310 6.89933730e-310 6.89933729e-310 6.89933730e-310]
 [6.89933730e-310 6.89933730e-310 6.89933729e-310 6.89933729e-310]
 [6.89933730e-310 6.89933729e-310 6.89933730e-310 6.89933729e-310]
 [6.89933730e-310 4.30513389e-317 4.30321296e-317 6.89933825e-310]
 [4.30389280e-317 6.89933822e-310 4.30366750e-317 6.89933822e-310]
 [4.30311810e-317 4.30480583e-317 4.30462401e-317 4.30336316e-317]
 [6.89933822e-310 4.30386513e-317 4.30358055e-317 4.30571886e-317]
 [4.30568724e-317 4.30659237e-317 6.89933822e-310 6.89933822e-310]
 [6.89933822e-310 6.89933822e-310 4.30289676e-317 6.89920336e-310]]

From shape or value


empty(shape[, dtype, order, like])	Return a new array of given shape and type, without initializing entries.
empty_like(prototype[, dtype, order, subok, …])	Return a new array with the same shape and type as a given array.
eye(N[, M, k, dtype, order, like])	Return a 2-D array with ones on the diagonal and zeros elsewhere.
identity(n[, dtype, like])	Return the identity array.
ones(shape[, dtype, order, like])	Return a new array of given shape and type, filled with ones.
ones_like(a[, dtype, order, subok, shape])	Return an array of ones with the same shape and type as a given array.
zeros(shape[, dtype, order, like])	Return a new array of given shape and type, filled with zeros.
zeros_like(a[, dtype, order, subok, shape])	Return an array of zeros with the same shape and type as a given array.
full(shape, fill_value[, dtype, order, like])	Return a new array of given shape and type, filled with fill_value.
full_like(a, fill_value[, dtype, order, …])	Return a full array with the same shape and type as a given array.

The source code is Open Source and can be found on GitHub.