求 NumPy 数组元素的和和乘积

在本文中，让我们讨论如何求 NumPy 数组的和与积。

NumPy 数组的总和

NumPy数组元素的求和可以通过以下方式实现

方法 #1：使用numpy.sum()

Syntax: numpy.sum(array_name, axis=None, dtype=None, out=None, keepdims=, initial=, where=)

例子：

Python3

# importing numpy
import numpy as np
 
 
def main():
 
    # initialising array
    print('Initialised array')
    gfg = np.array([[1, 2, 3], [4, 5, 6]])
    print(gfg)
     
    # sum along row
    print(np.sum(gfg, axis=1))
     
    # sum along column
    print(np.sum(gfg, axis=0))
     
    # sum of entire array
    print(np.sum(gfg))
     
    # use of out
    # initialise a array with same dimensions
    # of expected output to use OUT parameter
    b = np.array([0])  # np.int32)#.shape = 1
    print(np.sum(gfg, axis=1, out=b))
     
    # the output is stored in b
    print(b)
     
    # use of keepdim
    print('with axis parameter')
     
    # output array's dimension is same as specified
    # by the axis
    print(np.sum(gfg, axis=0, keepdims=True))
     
    # output consist of 3 columns
    print(np.sum(gfg, axis=1, keepdims=True))
     
    # output consist of 2 rows
    print('without axis parameter')
    print(np.sum(gfg, keepdims=True))
     
    # we added 100 to the actual result
    print('using initial parameter in sum function')
    print(np.sum(gfg, initial=100))
 
    # False allowed to skip sum operation on column 1 and 2
    # that's why output is 0 for them
    print('using where parameter ')
    print(np.sum(gfg, axis=0, where=[True, False, False]))
 
 
if __name__ == "__main__":
    main()

Python3

# importing numpy
import numpy as np
 
 
def main():
 
    # initialising array
    print('Initialised array')
    gfg = np.array([[1, 2, 3], [4, 5, 6]])
     
    print('original array')
    print(gfg)
     
    # cumulative sum of the array
    print(np.cumsum(gfg))
     
    # cumulative sum of the array along
    # axis 1
    print(np.cumsum(gfg, axis=1))
     
    # initialising a 2x3 shape array
    b = np.array([[None, None, None], [None, None, None]])
     
    # finding cumsum and storing it in array
    np.cumsum(gfg, axis=1, out=b)
     
    # printing resultant array
    print(b)
 
 
if __name__ == "__main__":
    main()

Python3

# importing numpy
import numpy as np
 
def main():
 
    # initialising array
    print('Initialised array')
    gfg = np.array([[1, 2, 3], [4, 5, 6]])
    print(gfg)
     
    # product along row
    print(np.prod(gfg, axis=1))
     
    # product along column
    print(np.prod(gfg, axis=0))
     
    # sum of entire array
    print(np.prod(gfg))
     
    # use of out
    # initialise a array with same dimensions
    # of expected output to use OUT parameter
    b = np.array([0])  # np.int32)#.shape = 1
    print(np.prod(gfg, axis=1, out=b))
     
    # the output is stored in b
    print(b)
     
    # use of keepdim
    print('with axis parameter')
     
    # output array's dimension is same as specified
    # by the axis
    print(np.prod(gfg, axis=0, keepdims=True))
     
    # output consist of 3 columns
    print(np.prod(gfg, axis=1, keepdims=True))
     
    # output consist of 2 rows
    print('without axis parameter')
    print(np.prod(gfg, keepdims=True))
     
    # we initialise product to a factor of 10
    # instead of 1
    print('using initial parameter in sum function')
    print(np.prod(gfg, initial=10))
     
    # False allowed to skip sum operation on column 1 and 2
    # that's why output is 1 which is default initial value
    print('using where parameter ')
    print(np.prod(gfg, axis=0, where=[True, False, False]))
     
if __name__ == "__main__":
    main()

Python3

# importing numpy
import numpy as np
 
 
def main():
 
    # initialising array
    print('Initialised array')
    gfg = np.array([[1, 2, 3], [4, 5, 6]])
    print('original array')
    print(gfg)
     
    # cumulative product of the array
    print(np.cumprod(gfg))
     
    # cumulative product of the array along
    # axis 1
    print(np.cumprod(gfg, axis=1))
     
    # initialising a 2x3 shape array
    b = np.array([[None, None, None], [None, None, None]])
     
    # finding cumprod and storing it in array
    np.cumprod(gfg, axis=1, out=b)
     
    # printing resultant array
    print(b)
 
 
if __name__ == "__main__":
    main()

输出：

Initialised array
[[1 2 3]
 [4 5 6]]
[ 6 15]
[5 7 9]
21
[21]
[21]
with axis parameter
[[5 7 9]]
[[ 6]
 [15]]
without axis parameter
[[21]]
using initial parameter in sum function
121
using where parameter 
[5 0 0]

注意：在包含非数字（NaNs）元素的数组元素上使用 numpy.sum 会产生错误，为避免这种情况，我们使用 numpy。 nansum()参数与前者类似，只是后者不支持where和initial。

方法 #2：使用numpy.cumsum()

返回给定数组中元素的累积和。

Syntax: numpy.cumsum(array_name, axis=None, dtype=None, out=None)

编程需要懂一点英语

例子：

Python3

# importing numpy
import numpy as np
 
 
def main():
 
    # initialising array
    print('Initialised array')
    gfg = np.array([[1, 2, 3], [4, 5, 6]])
     
    print('original array')
    print(gfg)
     
    # cumulative sum of the array
    print(np.cumsum(gfg))
     
    # cumulative sum of the array along
    # axis 1
    print(np.cumsum(gfg, axis=1))
     
    # initialising a 2x3 shape array
    b = np.array([[None, None, None], [None, None, None]])
     
    # finding cumsum and storing it in array
    np.cumsum(gfg, axis=1, out=b)
     
    # printing resultant array
    print(b)
 
 
if __name__ == "__main__":
    main()

输出：

Initialised array
original array
[[1 2 3]
 [4 5 6]]
[ 1  3  6 10 15 21]
[[ 1  3  6]
 [ 4  9 15]]
[[1 3 6]
 [4 9 15]]

NumPy 数组的乘积

NumPy 数组的乘积可以通过以下方式实现

方法 #1：使用numpy.prod()

Syntax: numpy.prod(array_name, axis=None, dtype=None, out=None, keepdims=, initial=, where=)

编程需要懂一点英语

例子：

Python3

# importing numpy
import numpy as np
 
def main():
 
    # initialising array
    print('Initialised array')
    gfg = np.array([[1, 2, 3], [4, 5, 6]])
    print(gfg)
     
    # product along row
    print(np.prod(gfg, axis=1))
     
    # product along column
    print(np.prod(gfg, axis=0))
     
    # sum of entire array
    print(np.prod(gfg))
     
    # use of out
    # initialise a array with same dimensions
    # of expected output to use OUT parameter
    b = np.array([0])  # np.int32)#.shape = 1
    print(np.prod(gfg, axis=1, out=b))
     
    # the output is stored in b
    print(b)
     
    # use of keepdim
    print('with axis parameter')
     
    # output array's dimension is same as specified
    # by the axis
    print(np.prod(gfg, axis=0, keepdims=True))
     
    # output consist of 3 columns
    print(np.prod(gfg, axis=1, keepdims=True))
     
    # output consist of 2 rows
    print('without axis parameter')
    print(np.prod(gfg, keepdims=True))
     
    # we initialise product to a factor of 10
    # instead of 1
    print('using initial parameter in sum function')
    print(np.prod(gfg, initial=10))
     
    # False allowed to skip sum operation on column 1 and 2
    # that's why output is 1 which is default initial value
    print('using where parameter ')
    print(np.prod(gfg, axis=0, where=[True, False, False]))
     
if __name__ == "__main__":
    main()

输出：

Initialised array
[[1 2 3]
 [4 5 6]]
[  6 120]
[ 4 10 18]
720
[720]
[720]
with axis parameter
[[ 4 10 18]]
[[  6]
 [120]]
without axis parameter
[[720]]
using initial parameter in sum function
7200
using where parameter 
[4 1 1]

方法 #2：使用numpy.cumprod()

返回数组的累积乘积。

Syntax: numpy.cumsum(array_name, axis=None, dtype=None, out=None)axis = [integer,Optional]

编程需要懂一点英语

Python3

# importing numpy
import numpy as np
 
 
def main():
 
    # initialising array
    print('Initialised array')
    gfg = np.array([[1, 2, 3], [4, 5, 6]])
    print('original array')
    print(gfg)
     
    # cumulative product of the array
    print(np.cumprod(gfg))
     
    # cumulative product of the array along
    # axis 1
    print(np.cumprod(gfg, axis=1))
     
    # initialising a 2x3 shape array
    b = np.array([[None, None, None], [None, None, None]])
     
    # finding cumprod and storing it in array
    np.cumprod(gfg, axis=1, out=b)
     
    # printing resultant array
    print(b)
 
 
if __name__ == "__main__":
    main()

输出：

Initialised array
original array
[[1 2 3]
 [4 5 6]]
[  1   2   6  24 120 720]
[[  1   2   6]
 [  4  20 120]]
[[1 2 6]
 [4 20 120]]