The following all use numpy's standard "import numpy as np" 1. numpy
is a multi-dimensional container for homogeneous data, and homogeneous means the same data type
2. Initialization:
2.1
np.arange([start,] end [, step])#与list的range相似
>>> np.arange(10)
array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
>>> np.arange(1, 10, 2)
array([1, 3, 5, 7, 9])
2.2
np.zeros(tupleA)# Generate a matrix of tupleA dimension, and the initial value is all 0
>>> np.zeros((4))
array([ 0., 0., 0., 0.])
>>> np.zeros((4,2))
array([[ 0., 0.],
[ 0., 0.],
[ 0., 0.],
[ 0., 0.]])
2.3
np.ones(tupleA)# is similar to the above, except that the initialization is all 1
>>> np.ones((4))
array([ 1., 1., 1., 1.])
>>> np.ones((4,2))
array([[ 1., 1.],
[ 1., 1.],
[ 1., 1.],
[ 1., 1.]])
2.4
np.empty(tupleA)# is similar to the above, but the initialization value is undefined (not 0 as you think!!!)
>>> np.empty((4))
array([ 1.73154357e-316, 4.71627160e-317, 0.00000000e+000,
4.94065646e-324])
>>> np.empty((3,2))
array([[ 0.00000000e+000, 0.00000000e+000],
[ 6.94647584e-310, 6.94647586e-310],
[ 6.94647586e-310, 6.94647586e-310],
2.5
np.array(listA)# Convert listA to np, listA is just a general term, as long as it is serialized, it can also be other np
>>> np.array([[1, 2, 3], [4, 3, 2]])
array([[1, 2, 3],
[4, 3, 2]])
>>> npA = np.array([[1, 2, 3], [4, 3, 2]])
>>> npA
array([[1, 2, 3],
[4, 3, 2]])
>>> npB = np.array([[1, 2, 3], [4, 3, 2.0]])
>>> npB
array([[ 1., 2., 3.],
[ 4., 3., 2.]])
np.array会自动找到最适合listA数据类型转给np:
>>> npA.dtype
dtype('int64')
>>> npB.dtype
dtype('float64')
But in fact, np will be defaulted to float64 without special instructions during initialization, such as the first four.
2.6
其他:ones_like(npA);zeros_like(npA);empty_like(npA)
>>> npB = np.array([[1, 2, 3], [4, 3, 2.0]])
>>> np.ones_like(npB)
array([[ 1., 1., 1.],
[ 1., 1., 1.]])
>>> np.zeros_like(npB)
array([[ 0., 0., 0.],
[ 0., 0., 0.]])
>>> np.empty_like(npB)
array([[ 0.00000000e+000, 0.00000000e+000, 1.56491143e-316],
[ 6.94647850e-310, 6.94635322e-310, 1.72361006e-316]])
>>> np.identity(3)
array([[ 1., 0., 0.],
[ 0., 1., 0.],
[ 0., 0., 1.]])
>>> np.eye(3, k = -1)#变化k的值试试看
array([[ 0., 0., 0.],
[ 1., 0., 0.],
[ 0., 1., 0.]])
Reprinted from: https://blog.csdn.net/u010668907/article/details/51429540