Source: blog.csdn.net/yuansuo0516…
C umsum: Computes the summation of axial elements and returns an array of intermediate results
The point is that the return value is”An array of intermediate results“
The following code runs successfully on python3.6!
Define an array of 2*2*3, so its shape is 2, 2, 3, and its indexes are 0,1,2
| shape | The index |
|---|---|
| 2 | 0 |
| 2 | 1 |
| 3 | 2 |
Code:
import numpy
as np
arr = np.array([[[
1.2.3], [8.9.12]], [[1.2.4], [2.4.5]]])
# 2 * 2 * 3
print(arr.cumsum(
0))
print(arr.cumsum(
1))
print(arr.cumsum(
2))
Copy the code
Copy the codeOutput result:#cumsum(0)
[[[ 1 2 3]
[ 8 9 12]]
[[ 2 4 7]
[10 13 17]]]
#cumsum(1)
[[[ 1 2 3]
[ 9 11 15]]
[[ 1 2 4]
[ 3 6 9]]]
#cumsum(2)
[[[ 1 3 6]
[ 8 17 29]]
[[ 1 3 7]
[ 2 6 11]]]
Copy the codeCopy the codeCopy the codeNote:
Arr is a 2*2*3 3d matrix with index values 0,1,2
Cumsum (0) : cumsum on axis 0: take the outermost element as the unit and start with [[1,2,3],[8,9,12]] to realize the corresponding accumulation of the following elements
Cumsum (1) : to realize the accumulation on axis 1: To realize the corresponding accumulation of the following elements starting from [1,2,3] with the middle array elements as the unit
Cumsum (2) : realize the accumulation on the 2 axis: start with the innermost element as the accumulation unit, i.e., 1, and realize the accumulation of the following elements
Four – dimensional array implementation
Let’s look at a four-dimensional array 2*2*2*4 with indexes 0,1,2,3
Code:
import numpy
as np
arr = np.arange(
32).reshape((
2.2.2.4))
print(arr)
print(arr.cumsum(
0))
print(arr.cumsum(
1))
print(arr.cumsum(
2))
print(arr.cumsum(
3))
Copy the code
arr:
[[[[
0
1
2
3]
[
4
5
6
7]]
[[
8
9
10
11]
[
12
13
14
15[[[]]]16
17
18
19]
[
20
21
22
23]]
[[
24
25
26
27]
[
28
29
30
31]]]]
Copy the codeArr is a2*2*2*4A four dimensional matrix with an index of0.1.2.3
Copy the codeCumsum (0) : implement the sum on the 0 axis namely: in the outermost element of the array namely
[[0 12 3] [4 5 6 7]] [[8 9 10 11] [12 13 14 15]]Copy the codewith
[[[
16
17
18
19]
[
20
21
22
23]]
[[
24
25
26
27]
[
28
29
30
31]]]]
Copy the codeAdd the elements in the corresponding position \
Results:
[[[[
0
1
2
3]
[
4
5
6
7]]
[[
8
9
10
11]
[
12
13
14
15[[[]]]16
18
20
22]
[
24
26
28
30]]
[[
32
34
36
38]
[
40
42
44
46]]]]
Copy the codeCumsum (1) : to realize the accumulation on axis 1, namely:
[0 1 2 3] [4 5 6 7]]Copy the codeadd
[[
8
9
10
11]
[
12
13
14
15]]
Copy the codeadd
[[16 17 18 19] [20 21 22 23]Copy the codeadd
[[
24
25
26
27]
[
28
29
30
31]]
Copy the codeThe intermediate results of the accumulative process are recorded in the array
So, as a result:
[[[[
0
1
2
3]
[
4
5
6
7]]
[[
8
10
12
14]
[
16
18
20
22[[[]]]16
17
18
19]
[
20
21
22
23]]
[[
40
42
44
46]
[
48
50
52
54]]]]
Copy the codeCumsum (2) starts from [0, 1, 2, 3] and records the intermediate result
Cumsum (3) = 0 cumsum(3) = 0 cumsum(3) = 0
Original explanation, prohibit copying, support reprint! Indicate the source **\
\