Exceptions

Errors in Python are called exceptions. For example, syntax error is a special type of exception where the code’s syntax is wrong, such as, forgetting : after if statement. You might even get errors when your code is syntactically correct. These are other types of exceptions.

Syntax Error Exceptions
No “:” after if statement Division by zero 2 / 0
Mismatched brackets [ ... ) Type errors 2 + "Hello"
Incorrect indentation Using an undefined object
… …

You have likely already seen some of these errors in the Python shell or in the code cells if your browser supports error messages.

E.g., entering print(2/0) in the Python shell will produce

>>> print(2 / 0)
Traceback (most recent call last):
  File "<console>", line 1, in <module>
ZeroDivisionError: division by zero

All error messages follow a similar format. Error message lists the location and a code snipped where the error happened together with the type of the error (e.g., ZeroDivisionError).

Do You Need to Memorise Errors?

You don’t have to memorise the exceptions and syntax errors. However, conceptual understanding of syntax rules, and common errors, and their messages will help you resolve most of the issues you face when writing code. You will naturally encounter and learn to recognise common errors as you gain more experience in Python.

Handling Exceptions

Errors disrupt the normal flow of the running program. When they occur the interpreter raises an exception. Python provides exception handling with try-except block. This allows managing the errors to prevent the program from crashing.

try:
    # code that might raise an exception
except ExceptionType as e:
    # catch ExceptionType errors
    # `e` represents the error

  • try block encloses the code that might raise exceptions, e.g., a code that reads files or a user input.

  • except block catches the exception that matches a given ExceptionType exception (error) type, e.g., ValueError, ZeroDivisionError, KeyError, etc.

  • You can use Exception type to cover all built-in exceptions except SystemExit.

SystemExit exception signals the Python interpreter to terminate. See the exceptions documentation to find out more about exceptions.

  • as keyword assigns the caught exception to the variable e.
  • You can also stack except blocks to handle multiple exceptions.
try:
    # code that might raise an exception
except ExceptionType as e:
    # catch ExceptionType errors
except AnotherExceptionType as e:
    # catch AnotherExceptionType

It is not necessary to call the error as e and you can use any valid name or omit the as e part altogether. If you omit the name, you need to refer to the exception using the class name of the exception (type, e.g., ZeroDivisionError).

Raising Exceptions

Python also allows you to force a specified exception with the raise statement.

raise ExceptionType

You can use raise with an exception with a custom message to force your own exceptions.

All exceptions accept an optional message string, ExceptionType(msg). The message is displayed when the error is encountered.

Forcing exception might seem counterintuitive. You might be wondering why would anyone need a way to raise errors. You have to keep in mind that not all programs run in isolation, and programs might interact with files, users, or communicate with other machines. In this circumstances a program is expected to encounter errors.

Assertions

Catching errors to gracefully handle, or intentionally raising them provides an additional way to control your program. This also can be used to explicitly incorporate programmer’s assumptions into the code. E.g., you might want to force a function to accept only a certain image size or format.

For convenience, Python also provides an assert statement in addition to the raise. Assertions are a convenient way to quickly include assumptions about a program in a single line of code.

assert condition

AssertionError is raised if the condition is False. You can also include a message string after the condition to describe your error.

assert condition, msg

E.g., in the below code cell div(a, b) runs only when b is non-zero.