[FIXED] Evaluating (False is False) is returning False in specific cases

Issue

I am encountering a strange problem in Python, where evaluating a condition involving a np.linalg.norm value is causing issues. This is my test script:

import numpy as np

def isGoalReached():
    start = np.array([4, 5, 6])
    goal = np.array([17, 18, 19])
    dist = np.linalg.norm(start - goal)
    return (dist < 0.5)

print (isGoalReached())
print (isGoalReached() == 0)
print (isGoalReached() is False)

Output:

False
True
False

If I remove the np.linalg.norm part, and rewrite so:

import numpy as np

def isGoalReached():
    dist = 123.456
    return (dist < 0.5)

print (isGoalReached())
print (isGoalReached() == 0)
print (isGoalReached() is False)

Output:

False
True
True

What is causing this very strange behavior? I am on Python 3.7.3 (Anaconda)

Solution

Seems np.linalg.norm returns a numpy.float64, and when you do the comparison, the type is numpy.bool_.

In [83]: np.bool_?                                                                                      
Init signature: np.bool_(self, /, *args, **kwargs)
Docstring:     
Boolean type (True or False), stored as a byte.

That is why is doesn't work. If you add

return bool(dist < 0.5)

it would work as you expect here.

Answered By - Wesley Bowman