Exceptions, Status Values and Tracebacks

Overview

Teaching: 10 min
Exercises: 0 min

Questions

• What does this error mean?

Objectives

• Learn how to read a Python traceback.

• Learn what an exception is, and that there are many types.

When a computer tries to perform some problematic task (e.g. try to run a python file with a typographic error, divide by zero), the computer needs to inform the program (and by extension the programmer and user) of the problem. In Python, this means either:

1. Returning a value to indicate an error occurred. This value is called a status value, a flag, or an error code. Status values you may have seen before include inf and NaN, and None is a commonly used status value in Python.
2. “Raising” an exception. The word “raising” in this context refers to how the exception is created and passed around by Python. Other languages may use the terminology “thrown”.

In general, Python programmers prefer using exceptions over status values, and we will provide some general guidance about when to use both in the next chapter.

You have likely seen quite a few different types of exceptions, for example the SyntaxError, which pops up when you’ve made a typographic error, or a NameError when you have misspelt a variable name. You may have also seen ZeroDivisionError when accidentally dividing by 0, or FileNotFoundError when you have misspelt a file name. All these different exceptions have two things in common: a message which tells you (sometimes cryptically) what went wrong, and where is occurred. Learning to debug Python is learning to use those two bits of information to find out why there is a problem.

What different types of exceptions are there?

A list of different exceptions included with Python can be found at https://docs.python.org/3/library/exceptions.html#exception-hierarchy. In addition, it is possible to create custom exceptions (which we won’t cover in this course).

Below are three Python functions which call each other in a Python file that I’ve called divide_by_zero.py.

def inner():
    return 1/0

def outer():
    return inner()

def more_outer():
    return outer()

more_outer()

and if we run the python file:

$ python divide_by_zero.py

we get the following output:

Traceback (most recent call last):
  File "divide_by_zero.py", line 10, in <module>
    more_outer()
  File "divide_by_zero.py", line 8, in more_outer
    return outer()
  File "divide_by_zero.py", line 5, in outer
    return inner()
  File "divide_by_zero.py", line 2, in inner
    return 1/0
ZeroDivisionError: integer division or modulo by zero

This output is called a traceback (or backtrace or stacktrace): the traceback contains the exception that was raised, where it was raised originally (line number, name of function and filename, where the function which called the function which raised the exception, and so on until we’re out of functions we called. Read tracebacks from bottom to top; the bottom is the location of the raised exception.

In the above example, the line

ZeroDivisionError: integer division or modulo by zero

tells us the exception name and a description of the problem (we tried to divide by zero); the lines

  File "divide_by_zero.py", line 2, in inner
    return 1/0

tell where the exception was raised (line 2 of divide_by_zero.py, inside the function called inner).

Whilst in our simple case we can directly see the problem, remember that just because an exception was raised on a particular line, it does not mean that the bug occurs on that line. Debugging is a skill which takes practice and experience to develop, and best way of improving it is to notice patterns in the types of bugs that you run into. Helping others with debugging their code is another way to gain experience.

Tools to use: Debuggers

Debuggers are a group of programs which allow users to step through the running of a program. Many text editors and programming environments provide either their own debugger, or build around a debugger. Python has a built in debugger pdb: usually it’s better to learn to use pdb than to add lots of print statements.

Key Points

• Exceptions are one way exposing errors.

• There are many types of exceptions.