Nov 8, 2017 6:00:31 PM | Python Exception Handling - FloatingPointError

A brief look at the FloatingPointError in Python, including functional code samples discussing and showing how such errors can be raised.

Today we get started with our in-depth Python Exception Handling series by looking at the FloatingPointError. As with most programming languages, the FloatingPointError in Python indicates that something has gone wrong with a floating point calculation. However, unlike most other languages, Python will not raise a FloatingPointError by default. The ability to do so must be implemented by including the fpectl module when building your local Python environment.

In this article we'll explore the FloatingPointError by first looking at where it resides in the overall Python Exception Class Hierarchy. We'll also go over how the fpectl module can be enabled, and how doing so can allow the raising of FloatingPointErrors in your own code. Let's get to it!

The Technical Rundown

All Python exceptions inherit from the BaseException class, or extend from an inherited class therein. The full exception hierarchy of this error is:

Full Code Sample

Below is the full code sample we'll be using in this article. It can be copied and pasted if you'd like to play with the code yourself and see how everything works.

import fpectl
from gw_utility.logging import Logging

def main():
Logging.line_separator("FLOATING POINT")
test_floating_point()

Logging.line_separator("DIVISION BY ZERO")
test_division_by_zero()

Logging.line_separator("FLOATING POINT DIVISION BY ZERO", 60)
test_floating_point_division_by_zero()

def test_floating_point():
try:
Logging.log(round(24.601 / 3.5, 4))
except FloatingPointError as exception:
# Output expected FloatingPointErrors.
Logging.log_exception(exception)
except Exception as exception:
# Output expected Exceptions.
Logging.log_exception(exception, False)

def test_division_by_zero():
try:
# Divide by zero.
Logging.log(24 / 0)
except FloatingPointError as exception:
# Output expected FloatingPointErrors.
Logging.log_exception(exception)
except Exception as exception:
# Output expected Exceptions.
Logging.log_exception(exception, False)

def test_floating_point_division_by_zero():
try:
# Divide by floating point zero and round.
Logging.log(round(24.601 / 0.0, 4))
except FloatingPointError as exception:
# Output expected FloatingPointErrors.
Logging.log_exception(exception)
except Exception as exception:
# Output expected Exceptions.
Logging.log_exception(exception, False)

if __name__ == "__main__":
main()

import math
import sys
import traceback

class Logging:

separator_character_default = '-'
separator_length_default = 40

@classmethod
def __output(cls, *args, sep: str=' ', end: str='\n', file=None):
"""Prints the passed value(s) to the console.

:param args: Values to output.
:param sep: String inserted between values, default a space.
:param end: String appended after the last value, default a newline.
:param file: A file-like object (stream); defaults to the current sys.stdout.
:return: None
"""
print(*args, sep=sep, end=end, file=file)

@classmethod
def line_separator(cls, value: str, length: int=separator_length_default, char: str=separator_character_default):
"""Print a line separator with inserted text centered in the middle.

:param value: Inserted text to be centered.
:param length: Total separator length.
:param char: Separator character.
"""
output = value

if len(value) < length:
# Update length based on insert length, less a space for margin.
length -= len(value) + 2
# Halve the length and floor left side.
left = math.floor(length / 2)
right = left
# If odd number, add dropped remainder to right side.
if length % 2 != 0:
right += 1

# Surround insert with separators.
output = f'{char * left} {value} {char * right}'

cls.__output(output)

@classmethod
def log(cls, *args, sep: str=' ', end: str='\n', file=None):
"""Prints the passed value(s) to the console.

:param args: Values to output.
:param sep: String inserted between values, default a space.
:param end: String appended after the last value, default a newline.
:param file: A file-like object (stream); defaults to the current sys.stdout.
"""
cls.__output(*args, sep=sep, end=end, file=file)

@classmethod
def log_exception(cls, exception: BaseException, expected: bool=True):
"""Prints the passed BaseException to the console, including traceback.

:param exception: The BaseException to output.
:param expected: Determines if BaseException was expected.
"""
output = "[{}] {}: {}".format('EXPECTED' if expected else 'UNEXPECTED', type(exception).__name__, exception)
cls.__output(output)
exc_type, exc_value, exc_traceback = sys.exc_info()
traceback.print_tb(exc_traceback)

When Should You Use It?

As discussed in the introduction, before a FloatingPointError can even appear you'll need to make sure your local Python build includes the fpectl module. Since this module is not included with most Python builds by default, you'd likely have had to explicitly build your Python with it if desired. Adding the fpectl module to can be accomplished by using the --with-fpectl flag when compiling Python. Going through the compilation process of Python is well beyond the scope of this article, but once fpectl is an included module, you can start testing the FloatingPointError.

For our example code we're not doing anything spectacular. In fact, the FloatingPointError is effectively raised in situations where other ArithmeticErrors would normally appear, except that you're using floating point numbers and the fpectl module is enabled. For example, you might raise a FloatingPointError where you'd normally get a ZeroDivisionError by attempting to divide by zero using a floating point value.

We've created a few simple testing methods starting with test_floating_point():

def test_floating_point():
try:
Logging.log(round(24.601 / 3.5, 4))
except FloatingPointError as exception:
# Output expected FloatingPointErrors.
Logging.log_exception(exception)
except Exception as exception:
# Output expected Exceptions.
Logging.log_exception(exception, False)

Executing this code works as expected, performing the floating point calculation and rounding the result to four decimal places before outputting the result to our log:

------------ FLOATING POINT ------------
7.0289

Now, let's step away from using a floating point value and use regular integers while attempting to divide by zero:

def test_division_by_zero():
try:
# Divide by zero.
Logging.log(24 / 0)
except FloatingPointError as exception:
# Output expected FloatingPointErrors.
Logging.log_exception(exception)
except Exception as exception:
# Output expected Exceptions.
Logging.log_exception(exception, False)

This raises an unexpected ZeroDivisionException since, even though fpectl is enabled, we aren't using a floating point value in our calculation:

----------- DIVISION BY ZERO -----------
[UNEXPECTED] ZeroDivisionError: division by zero
File "D:/work/Airbrake.io/Exceptions/Python/BaseException/Exception/ArithmeticError/FloatingPointError/main.py", line 30, in test_division_by_zero
Logging.log(24 / 0)

Finally, let's try the same division by zero while using floating point values:

def test_floating_point_division_by_zero():
try:
# Divide by floating point zero and round.
Logging.log(round(24.601 / 0.0, 4))
except FloatingPointError as exception:
# Output expected FloatingPointErrors.
Logging.log_exception(exception)
except Exception as exception:
# Output expected Exceptions.
Logging.log_exception(exception, False)

As you might suspect, this raises a FloatingPointError for us:

------------- FLOATING POINT DIVISION BY ZERO --------------
[EXPECTED] FloatingPointError: invalid value encountered in divide
File "D:/work/Airbrake.io/Exceptions/Python/BaseException/Exception/ArithmeticError/FloatingPointError/main.py", line 42, in test_floating_point_division_by_zero
Logging.log(round(24.601 / 0.0, 4))

There we have the basics of using FloatingPointErrors. However, before you jump into adding the fpectl module to your Python to distinguish between FloatingPointErrors and normal ArithmeticErrors, there are a number of caveats and cautions to be aware of. The IEEE 754 standard for floating point arithmetic defines a number of universal standards for the formatting, rounding, allowed operations, and exception handling practices of floating point numbers. However, your code must be explicitly told to capture IEEE 754 exceptions in the form of SIGFPE signals generated by the local processor. Consequently, while Python is configured to do so via the fpectl module, many other custom scripts/applications are not.

The other major consideration is that use of the fpectl module is generally discouraged, in large part because it is not thread safe. Thread safe applications (that is, most properly developed Python applications) allow data structures to be safely shared between multiple threads without fear of one thread manipulating or altering some data that another thread is using (or where another thread sees different data). However, using the fpectl module means your floating point data is no longer thread safe, which could cause major issues in multithreaded applications. To be on the safe side, it's generally recommended that you avoid fpectl and use another form of application logic to check for arithmetic errors.

Airbrake's robust error monitoring software provides real-time error monitoring and automatic exception reporting for all your development projects. Airbrake's state of the art web dashboard ensures you receive round-the-clock status updates on your application's health and error rates. No matter what you're working on, Airbrake easily integrates with all the most popular languages and frameworks. Plus, Airbrake makes it easy to customize exception parameters, while giving you complete control of the active error filter system, so you only gather the errors that matter most.

Check out Airbrake's error monitoring software today and see for yourself why so many of the world's best engineering teams use Airbrake to revolutionize their exception handling practices! Try Airbrake free with a 14-day trial.

Written By: Frances Banks