Python Structured arrays instead of many single arrays

Question

I am new to structured arrays, but I think it is that what I need.

I have a numpy array which contains for every entry a minus counter and one array that contains a plus counter. They are filled within a process. At the end of this process I want to divide all plus counter cell per cell through the minus counter and save it in an array with the result of this calculation in every entry.

So I think there is a better way to create three different arrays with the same size. Lets assume my arrays are in this form:

import numpy as np
import itertools 

plusArr = np.ones((20 ,39, 90))
minusArr = np.ones((20 ,39, 90))
resultArr = np.zeros((20 ,39, 90))

Then the cells are filled with numbers. At the end I doing something like:

for i in itertools.product(np.arange(0,20,1),np.arange(0,39,1),np.arange(0,90,1)):
    resultArr[i[0]][i[1]][i[2]] = plusArr[i[0]][i[1]][i[2]]/minusArr[i[0]][i[1]][i[2]]
print(resultArr)

This works but in the filling process it is very time consuming to look up the same entry location for both, minus and plus array. So I thought maybe to have a structured array with a triple instead of i int entries. In every triple the first entry is the plus counter, the second is the minus counter, and the last entry is 0 until the filling process is finished and it can be filled with the result of the division of plus and minus entry.

As a small side question. If one of my counter entries is 0, then i want to fill the result cell of the entry of the array which has no zero entry at this location. Is there a smart numpy way to do that or just an if condition?

The answer looks good, but to be sure you should give us a small example - with real values and what you expect. No one wants to display or test (20,39,90) size arrays. Also I don't see why you thing structured arrays will help you. For math you have to treat the fields as separate arrays, which you already have. — hpaulj
– hpaulj, Commented Jan 15, 2018 at 18:53

Stephen Rauch · Accepted Answer · 2018-01-15 18:01:23Z

1

Numpy allows basic element by element arithmetic operations directly on the arrays. In addition there is whole bunch of math routines available.

Code:

The element by element division is as simple as:

result_arr = plus_arr / minus_arr

Test Code:

import numpy as np

size = (1, 2, 3)
plus_arr = np.ones(size)
minus_arr = np.ones(size) * 2
minus_arr[0, 0, 0] = 3

result_arr = plus_arr / minus_arr

print(plus_arr)
print(minus_arr)
print(result_arr)

Test Results:

[[[ 1.  1.  1.]
  [ 1.  1.  1.]]]

[[[ 3.  2.  2.]
  [ 2.  2.  2.]]]

[[[ 0.33333333  0.5         0.5       ]
  [ 0.5         0.5         0.5       ]]]

answered Jan 15, 2018 at 18:01

Stephen Rauch♦

50.1k32 gold badges118 silver badges143 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Python Structured arrays instead of many single arrays

1 Answer 1

Code:

Test Code:

Test Results:

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Code:

Test Code:

Test Results:

Comments

Your Answer

Sign up or log in

Post as a guest

Related