Numpy starter tutorial in Python
- 2020-04-02 13:35:27
- OfStack
1. What is Numpy
Quite simply, Numpy is a library for scientific computing in Python, providing the functionality of matrix operations, which is commonly used in conjunction with Scipy and matplotlib. List already provides a matrix-like representation, but numpy gives us more functions. If you have been exposed to matlab, scilab, then numpy is a good start. In the following code example, numpy is always preloaded:
>>> import numpy as np
>>> print np.version.version
1.6.2
2. Multidimensional array
The type of the multidimensional array is: numpy.ndarray.
Use the numpy.array method
A one-dimensional array is generated by taking a list or tuple variable as a parameter:
>>> print np.array([1,2,3,4])
[1 2 3 4]
>>> print np.array((1.2,2,3,4))
[ 1.2 2. 3. 4. ]
>>> print type(np.array((1.2,2,3,4)))
<type 'numpy.ndarray'>
To generate a two-dimensional array with a list or tuple variable as an element:
>>> print np.array([[1,2],[3,4]])
[[1 2]
[3 4]]
When an array is generated, you can specify the data types, such as numpy.int32, numpy.int16, and numpy.float64:
>>> print np.array((1.2,2,3,4), dtype=np.int32)
[1 2 3 4]
Use the numpy.arange method
>>> print np.arange(15)
[ 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14]
>>> print type(np.arange(15))
<type 'numpy.ndarray'>
>>> print np.arange(15).reshape(3,5)
[[ 0 1 2 3 4]
[ 5 6 7 8 9]
[10 11 12 13 14]]
>>> print type(np.arange(15).reshape(3,5))
<type 'numpy.ndarray'>
Use the numpy.linspace method
For example, generate 9 Numbers from 1 to 3:
>>> print np.linspace(1,3,9)
[ 1. 1.25 1.5 1.75 2. 2.25 2.5 2.75 3. ]
Specific matrices can be constructed using methods such as numpy.zeros, numpy.ones, and numpy.eye
Such as:
>>> print np.zeros((3,4))
[[ 0. 0. 0. 0.]
[ 0. 0. 0. 0.]
[ 0. 0. 0. 0.]]
>>> print np.ones((3,4))
[[ 1. 1. 1. 1.]
[ 1. 1. 1. 1.]
[ 1. 1. 1. 1.]]
>>> print np.eye(3)
[[ 1. 0. 0.]
[ 0. 1. 0.]
[ 0. 0. 1.]]
Create a 3d array:
>>> print np.zeros((2,2,2))
[[[ 0. 0.]
[ 0. 0.]]
[[ 0. 0.]
[ 0. 0.]]]
Get array properties:
>>> a = np.zeros((2,2,2))
>>> print a.ndim # The number of dimensions of an array
3
>>> print a.shape # The size of each dimension of the array
(2, 2, 2)
>>> print a.size # The number of elements in an array
8
>>> print a.dtype # The element type
float64
>>> print a.itemsize # The number of bytes per element
8
Array index, slice, assign
Example:
>>> a = np.array( [[2,3,4],[5,6,7]] )
>>> print a
[[2 3 4]
[5 6 7]]
>>> print a[1,2]
7
>>> print a[1,:]
[5 6 7]
>>> print a[1,1:2]
[6]
>>> a[1,:] = [8,9,10]
>>> print a
[[ 2 3 4]
[ 8 9 10]]
Use the for action element
>>> for x in np.linspace(1,3,3):
... print x
...
1.0
2.0
3.0
Basic array operations
First, construct arrays a and b:
>>> a = np.ones((2,2))
>>> b = np.eye(2)
>>> print a
[[ 1. 1.]
[ 1. 1.]]
>>> print b
[[ 1. 0.]
[ 0. 1.]]
Array addition, subtraction, multiplication and division:
>>> print a > 2
[[False False]
[False False]]
>>> print a+b
[[ 2. 1.]
[ 1. 2.]]
>>> print a-b
[[ 0. 1.]
[ 1. 0.]]
>>> print b*2
[[ 2. 0.]
[ 0. 2.]]
>>> print (a*2)*(b*2)
[[ 4. 0.]
[ 0. 4.]]
>>> print b/(a*2)
[[ 0.5 0. ]
[ 0. 0.5]]
>>> print (a*2)**4
[[ 16. 16.]
[ 16. 16.]]
Using methods that come with array objects:
>>> a.sum()
4.0
>>> a.sum(axis=0) # Calculate the sum of each column (a matrix-like column in a two-dimensional array)
array([ 2., 2.])
>>> a.min()
1.0
>>> a.max()
1.0
Using the method under numpy:
>>> np.sin(a)
array([[ 0.84147098, 0.84147098],
[ 0.84147098, 0.84147098]])
>>> np.max(a)
1.0
>>> np.floor(a)
array([[ 1., 1.],
[ 1., 1.]])
>>> np.exp(a)
array([[ 2.71828183, 2.71828183],
[ 2.71828183, 2.71828183]])
>>> np.dot(a,a) ## Matrix multiplication
array([[ 2., 2.],
[ 2., 2.]])
Merge array
Use the vstack and hstack functions under numpy:
>>> a = np.ones((2,2))
>>> b = np.eye(2)
>>> print np.vstack((a,b))
[[ 1. 1.]
[ 1. 1.]
[ 1. 0.]
[ 0. 1.]]
>>> print np.hstack((a,b))
[[ 1. 1. 1. 0.]
[ 1. 1. 0. 1.]]
Let's see if these two functions involve shallow copying:
>>> c = np.hstack((a,b))
>>> print c
[[ 1. 1. 1. 0.]
[ 1. 1. 0. 1.]]
>>> a[1,1] = 5
>>> b[1,1] = 5
>>> print c
[[ 1. 1. 1. 0.]
[ 1. 1. 0. 1.]]
As you can see, the change of elements in a and b does not affect c.
Deep copy array
Array objects come with shallow-copy and deep-copy methods, but there are usually more deep-copy methods:
>>> a = np.ones((2,2))
>>> b = a
>>> b is a
True
>>> c = a.copy() # Deep copy
>>> c is a
False
Basic matrix operations
Transpose:
>>> a = np.array([[1,0],[2,3]])
>>> print a
[[1 0]
[2 3]]
>>> print a.transpose()
[[1 2]
[0 3]]
Mark:
>>> print np.trace(a)
4
The numpy.linalg module has a number of methods for matrix operations:
>>> import numpy.linalg as nplg
Eigenvalues and eigenvectors:
>>> print nplg.eig(a)
(array([ 3., 1.]), array([[ 0. , 0.70710678],
[ 1. , -0.70710678]]))
3, matrix
Numpy can also construct matrix objects, which I won't discuss here.