This article is to interpret some The commonly used numpy array operation is also to deepen your understanding of numpy. If you encounter more practical operations later, you will also synchronize to this article.
It has to be said that numpy has too many operations. A good way is to use it while learning, use it at any time and check it at any time. If you want to learn comprehensively or just want to query the function of a function, you can Visit the official website , the official website has a very comprehensive tutorial method, which is a great place to learn numpy.
1. Calculation of array and scalar
| Operator operation | meaning |
|---|---|
| a+1 | Add 1 to all elements of a |
| a - 1 | All elements of a minus 1 |
| a*3 | Multiply all elements of a by 3 |
| 1/a | 1 divided by all elements in a |
| 1//a | 1 divided by all elements in a and rounded |
| a**0.5 | All elements of a are taken to the power of 0.5 |
Code example:
import numpy as np a = np.array([1,2,3,4]).reshape(2,2) print(a+2) #addition ''' [[3 4] [5 6]] ''' print(a*3) #multiplication ''' [[ 3 6] [ 9 12]] ''' print(1//a) # Division ''' [[1 0] [0 0]] ''' print(a**0.5) #Open root ''' [[1. 1.41421356] [1.73205081 2. ]] '''
2. Mathematical calculation between two arrays
| Operator operation | meaning |
|---|---|
| a+b | a. B addition of corresponding elements |
| a-b | a. B subtraction of corresponding elements |
| a*b | a. B multiplies the corresponding elements |
| a/b | a. B Division of corresponding elements |
| a//b | a. B Division of corresponding elements |
| np.multiply(a, b) | Dot multiplication, same as * sign |
| np.dot(a,b) | inner product |
| np.matmul | Cross multiplication |
Code example:
import numpy as np a = np.array([1.2,2.6,-3.9,4.5]).reshape(2,2) b = np.array([-2.6,3.6,2.8,13.6]).reshape(2,2) print(a*b) ''' [[ -3.12 9.36] [-10.92 61.2 ]] ''' print(a//b) ''' [[-1. 0.] [-2. 0.]] '''
3. Mathematical operation of array unary function
| function | meaning |
|---|---|
| numpy.sqrt(array) | Square root function |
| numpy.exp(array) | Array of e^array[i] |
| numpy.abs/fabs(array) | Calculate absolute value |
| numpy.square(array) | Calculate the square of each element, equal to array**2 |
| numpy.log/log10/log2(array) | Calculate various logarithms of each element |
| numpy.sign(array) | Calculate the sign of each element |
| numpy.isnan(array) | NaN(not a number): not equal to any floating-point number |
| numpy.isinf(array) | inf: larger than any floating point number |
| numpy.cos/cosh/sin/sinh/tan/tanh(array) | trigonometric function |
| numpy.modf(array) | The integer and decimal values in the array are separated and returned as two arrays |
| numpy.ceil(array) | Round up, that is, take an integer larger than this number |
| numpy.floor(array) | Round down, that is, take an integer smaller than this number |
| numpy.rint(array) | rounding |
| numpy.trunc(array) | Round to 0 |
| numpy.cos(array) | sine |
| numpy.sin(array) | Cosine value |
| numpy.tan(array) | Tangent value |
| numpy.sum(array) | Sum |
| numpy.cumsum(array) | Prefix Sum |
| numpy.mean(array) | Average |
| numpy.std(array) | Standard deviation |
| numpy.var(array) | Find variance |
| numpy.min(array) | Find the minimum value |
| numpy.max(array) | Find the maximum value |
| numpy.argmin(array) | Minimum index |
| numpy.argmax(array) | Maximum index |
| numpy.media(array) | The odd number takes the middle number, and the even number takes the average of the middle two numbers |
Code example:
import numpy as np a = np.array([1.2,2.6,-3.9,4.5]).reshape(2,2) print(np.rint(a)) #rounding ''' [[ 1. 3.] [-4. 4.]] ''' print(np.abs(a)) #absolute value ''' [[1.2 2.6] [3.9 4.5]] ''' print(np.sin(a)) #sin value ''' [[ 0.93203909 0.51550137] [ 0.68776616 -0.97753012]] ''' print(np.sum(a)) #Summation 4.4 print(np.cumsum(a)) # Sum of current number and all previous numbers: [1.2 3.8 - 0.1 4.4] print(np.argmin(a)) #Minimum index: 2
4. Mathematical operation of array binary function
| function | meaning |
|---|---|
| numpy.add(array1,array2) | Element level addition |
| numpy.subtract(array1,array2) | Element level subtraction |
| numpy.multiply(array1,array2) | Element level multiplication |
| numpy.divide(array1,array2) | Element level division array1/ array2 |
| numpy.power(array1,array2) | Element level index array1^ array2 |
| numpy.maximum/minimum(array1,aray2) | Element level maximum |
| numpy.fmax/fmin(array1,array2) | Element level maximum, ignoring NaN |
| numpy.mod(array1,array2) | Element level modulus |
| numpy.copysign(array1,array2) | Copies the value symbol from the second array to the value in the first array |
| numpy.greater/greater_equal/less/less_equal/equal/not_equal (array1,array2) | Element level comparison operation to generate boolean array |
| numpy.logical_end/logical_or/logic_xor(array1,array2) | Element level truth logic operation |
Code example:
import numpy as np a = np.array([1.2,2.6,-3.9,4.5]).reshape(2,2) b = np.array([-2.6,3.6,2.8,13.6]).reshape(2,2) print(np.add(a,b)) #Addition of elements ''' [[-1.4 6.2] [-1.1 18.1]] ''' print(np.multiply(a,b)) #Element multiplication ''' [[ -3.12 9.36] [-10.92 61.2 ]] ''' print(np.maximum(a,b)) #a. Maximum element between B ''' [[ 1.2 3.6] [ 2.8 13.6]] ''' print(np.copysign(a,b))#Give the symbol of b to a ''' [[-1.2 2.6] [ 3.9 4.5]] '''
5. Common attributes of array
| attribute | meaning |
|---|---|
| dtype | The data type of the array element |
| size | Number of array elements |
| ndim | Dimension of array |
| shape | Dimension size of the array (in tuples) |
| T | Transpose of array |
Code example:
import numpy as np a = np.array([1,2,3,4]).reshape(2,2) print(a.dtype) #Type: int64 print(a.ndim)#Dimension: 2 print(a.shape)#Array size (2,2) print(a.size)#Number of array elements: 4 print(a.T) #Transpose ''' [[1 3] [2 4]] '''
6. Common methods of array
Note that this is a method, not a function.
| method | meaning |
|---|---|
| reshape | Modify shape without changing data |
| flat | Not a method, it's an array element iterator |
| flatten | Returns a copy of the array. Changes made to the copy will not affect the original array |
| ravel | Returns an expanded array |
Both flatten and t ravel can select the expansion order: Order: 'C' - by row, 'F' - by column, 'A' - original order, 'k' - the order of elements in memory.
import numpy as np
a = np.arange(4).reshape(2, 2)
for element in a.flat:
print(element,end=", ") # 0, 1, 2, 3,
print('\n')
print (a.flatten()) #[0 1 2 3]
print (a.ravel(order = 'F')) #[0 2 1 3]
7. Random operation
numpy. The random module complements Python's built-in random and adds some functions for efficiently generating sample values of multiple probability distributions, such as normal distribution, Poisson distribution, etc. Since the application of using normal distribution has not been involved yet, this part only introduces some simple and practical random operations.
| operation | significance |
|---|---|
| random.rand() | Generate data between [0,1] according to the given dimension |
| random.randn() | According to a given dimension (a single number when no dimension is given), a random number conforming to the standard normal distribution is generated |
| random.normal() | A random number that can define the normal distribution of mean and standard deviation |
| random.randint() | Returns a random integer for a given dimension |
| random.random(),random.random_sample() | Returns a random number between [0,1] of a given dimension |
| random.choice() | Generates a random number from a given one-dimensional array |
Code example:
import numpy as np print(np.random.rand(3, 3)) #Random number of [0,1) in 3 * 3 dimension print(np.random.random((3, 3))) #Random number of [0,1) in 3 * 3 dimension print(np.random.randint(0, 2, size=(3, 3))) #Random integer of [0, 2] of 3 * 3 dimension print(np.random.choice(['a', 'b', 'c'], p=[0.5, 0.3, 0.2])) #The sum of the probability values of p must be 1
8. Splicing array
import numpy as np a = np.arange(4).reshape(2,2) b = np.array([[5,6],[7,8]]) c = np.concatenate((a,b),axis=1)#Specify axis=1 splicing, and the default is axis=0 splicing ''' [[0 1 5 6] [2 3 7 8]] ''' c = np.concatenate((a,b))#Default axis=0 ''' [[0 1] [2 3] [5 6] [7 8]] ''' c=np.hstack([a,b])#Transverse splicing ''' [[0 1 5 6] [2 3 7 8]] ''' c=np.vstack([a,b])#Longitudinal splicing ''' [[0 1] [2 3] [5 6] [7 8]] ''' b = np.array([[1,1],[1,1]]) c = np.append(b,a)#If axis is not specified, it will be expanded into one dimension by default. If axis is specified, it is as follows. ''' [1 1 1 1 0 1 2 3 4 5 6 7] '''
axis is set in append.
import numpy as np
a=[1,2,3]
b=[4,5]
c=[[6,7],[8,9]]
d=[[10,11],[12,13]]
print('In one-dimensional array a Add after values,The results are as follows:{0}'.format(np.append(a,b,axis=0)))
print('Along 2D array c Add new row direction values The results are as follows:{0}'.format(np.append(c,d,axis=0)))
print('Along 2D array c Column direction add values The results are as follows:{0}'.format(np.append(c,d,axis=1)))
#print('axis is used. If the shapes of arr and values are different, an error will be reported: '. format(np.append(a,c,axis=0)))
Output:
In one-dimensional array a Add after values,The results are as follows:[1 2 3 4 5] Along 2D array c Add new row direction values The results are as follows:[[ 6 7] [ 8 9] [10 11] [12 13]] Along 2D array c Column direction add values The results are as follows:[[ 6 7 10 11] [ 8 9 12 13]]
9. Split array
import numpy as np a = np.arange(9) print(np.split(a,[3,6]))#The original array is segmented according to the specified segmentation point ''' [array([0, 1, 2]), array([3, 4, 5]), array([6, 7, 8])] '''
10. Insertion and deletion
import numpy as np a = np.arange(8).reshape(2,4) c = np.delete(a,[0,2],axis=1)#Delete on axis = 1, column 0,2 ''' [[1 3] [5 7]] ''' c = np.insert(a,1,5,axis=1)#Insert the specified value on axis=1, and insert the number 5 in column 1 ''' [[0 5 1 2 3] [4 5 5 6 7]] '''
11. Array extension
numpy provides tile function and repeat method, which can easily expand the array.
import numpy as np a = np.array([1,2,3]) print(np.tile(a,[3,1])) #Copy into 3 rows and 1 column ''' [[1 2 3] [1 2 3] [1 2 3]] ''' print(a.repeat(3,axis=0))#Copy 3 lines above axis=0 ''' [1 1 1 2 2 2 3 3 3] ''' a=np.array([[1,2,3]]) print(np.tile(a,[3,1])) #Copy into 3 rows and 1 column ''' [[1 2 3] [1 2 3] [1 2 3]] ''' print(a.repeat(3,axis=0))#Copy 3 lines above axis=0 ''' [[1 2 3] [1 2 3] [1 2 3]] '''