How to raise error for nan value in python numpy

Question

I need to automatically raise an error, if I divide by zero in numpy array using the "/" operator, or better if I do any operation, that results in nan.

What numpy does instead is to return nan value. I do not need my program to run after that, since there is obviously a mistake.

I want a way to know when division by zero is used and on which line. Just forbid any value of nan to be even created.

Answer 1

You can do some thing like this:

# sample data
arr1 = np.array([1,2,3,4])
arr2 = np.array([1,1,0,1])

# this will raise the error
assert (arr2!=0).all(), 'Division by zero'

# divide
arr1/arr2

And when you run the code, you would get:

      4 
      5 # this will raise the error
----> 6 assert (arr2!=0).all(), 'Division by zero'
      7 
      8 # divide

AssertionError: Division by zero

Answer 2

The default behavior is to emit a warning, but you can set the behavior using np.seterr:

>>> import numpy as np
>>> np.array([1,2,3]) / np.array([0, 1, 2])
__main__:1: RuntimeWarning: divide by zero encountered in true_divide
array([inf, 2. , 1.5])

After we change it to "raise":

>>> np.seterr(divide='raise')
{'divide': 'warn', 'over': 'warn', 'under': 'ignore', 'invalid': 'warn'}
>>> np.array([1,2,3]) / np.array([0, 1, 2])
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
FloatingPointError: divide by zero encountered in true_divide