Make your Python code more robust
Photo by ?? ? on Unsplash
When we learn coding in Python, we inevitably make various mistakes, most of the time syntactically and sometimes semantically. For example, I don?t know how many times I have forgotten the colon following an if statement or a function declaration, which results in the followingSyntaxError:
Syntax Error
These syntax errors, also known as parsing errors, are usually indicated by a little upward arrow in Python, as shown in the code snippet above. Besides paring errors, our code can contain other mistakes that are of more logical problems. For example, the TypeError is another error message we frequently encounter:
Exception Example
In the above code snippet, we were trying to concatenate strings. However, the number 2020 is of the type int, which can?t be used in a string concatenation that works with only str objects. In this case, we?ll encounter the TypeError. Using more precise jargon, the TypeError exception is raised or Python raises the TypeError exception.
Like TypeError, these kinds of errors (e.g., ValueError and ZeroDivisionError) happen when Python is trying to execute these lines of code. In Python, we typically term these non-syntax errors as exceptions, making them distinct from syntax errors. For syntax errors, we have to update the affected lines of code by using the acceptable syntax. For exceptions, we can handle them gracefully with the proper implementation of relevant techniques.
Let?s first take a look at how we can handle exceptions. In the second half, we?ll learn about exception-raising in Python.
Exception Handling
1. Basic form of handling exceptions
The standard way to handle exceptions is to use the try?except block. It?s pretty much like try?catch block in many other programming languages, if you have such a background.
The try clause includes the code that potentially raises an exception. If everything works well in the try clause, no code in the except clause will be executed. The try?except block is completed and the program will proceed.
However, if an exception is raised in the try clause, Python will stop executing any more code in that clause, and pass the exception to the except clause to see if this particular error is handled there.
Let?s take a look at a trivial example of the most basic form of exception handling:
Basic Form of Exception Handling
As you can see, when the division works as expected, the result of this division (i.e., 2.0) is printed. However, when we try to divide the number by zero, Python raises the ZeroDivisionError. Fortunately, our function was written to handle this error, and the message ?You can?t divide 12 by zero.? is printed to inform the user of this error.
2. Why handle exceptions?
We now understand how to handle exceptions using the try?except block. But why do we bother to handle exceptions? The most essential benefit is to inform the user of the error, while still allowing the program to proceed. Let?s see some similar functions, with and without handling exceptions:
Exception Handled vs. Un-handled
As shown above, when we call the function that handles the exception, we see that the program executes until the end of the function (Lines 15?17). By contrast, when we call the function that doesn?t handle the exception, we see that the program can?t complete to the end of the function (Lines 18?22).
3. Variable assignment
We can assign the exception to a variable such that we can retrieve more information about the exception. In the code below, we can assign the handled exception TypeError to the variable e, so we can ask Python to print the error message for us. As shown in Line 10, the error message is printed telling us that we can?t concatenate strings with integers:
Variable Assignment of Exception
4. Multiple exceptions
We can handle multiple exceptions in the except clause. We?ll simply wrap possible exceptions in a tuple, as shown in Line 6 in the following code snippet. When we call the function, we intentionally make two distinct errors by raising the ValueError and ZeroDivisionError, respectively. The messages clearly tell us what exceptions are handled.
Multiple Exceptions
5. Multiple except clauses
Related to the previous section, when we expect different exceptions, we can actually have multiple except clauses with each handling some specific exceptions. Let?s modify the above function (i.e., divide_six) to create multiple except clauses, as shown below.
Multiple Except Clauses
After the modification, when we call the function twice with the intention of raising two distinct exceptions each, the expected messages are printed for each except clause.
6. The else clause
We can use an else clause in the try?except block. It should be noted that the else clause needs to appear after the except clause. The code in the else clause runs when the try clause completes without any exceptions raised. Let?s see it in use:
Exception Handling With The Else Clause
The code has a function that uses an else clause in the try?except block. As you can see, the code in the else clause only runs when the try clause completes and no exceptions are raised. On the other hand, the code does not run when an exception is raised and handled.
7. The finally clause
Besides the use of the else clause, we can also use a finally clause in the try?except block. Please note that the finally clause needs to be placed at the end of the block, below the except clause or else clause (if set). The code in the finally clause will run right before the entire try?except block completes (after executing code in the try or except clause). Importantly, the code in the finally clause will run regardless of the exception raising and handling status. Let?s see how it works:
Exception Handling With The Finally Clause
As shown in the code snippet above, we have a function that has a finally clause. We call the function twice with the second call raising an exception. In both cases, the code in the finally clause runs successfully.
Another important thing to note with the use of the finally clause is that if the try clause includes a break, continue, and return statement, the finally clause will run first before executing the break, continue, or return statement. Many people can make mistakes here. Let?s take a look at a trivial example below:
Order of Execution (finally)
Raise Exceptions
8. Basic form of raising exceptions
In the last section, we learned various features of using the try?except block to handle exceptions in Python, which are certainly necessary for more robust code. However, your code can be further strengthened if you know how to raise exceptions properly. Let?s first see a basic form:
Raising Exception: Basic Form
As shown above, we use the raise keyword (in other programming languages, it?s called throw), followed by the exception class (e.g., Exception, NameError). We can also use the exception class constructor to create an instance, like ValueError(). These two usages have no differences, and the former is just a syntax sugar for the latter using the constructor.
9. Exception with a custom message
We can also provide additional information about the exception that we?re raising. The easiest way to do it is simply to use the exception class constructor and include the applicable error message to create the instance. This is pretty straightforward:
Custom Message
10. Re-raise and bubble up
As shown, all the exceptions are handled whenever they?re caught. However, it?s possible that we can re-raise the exception and pass the exception to the outside scope to see if it can be handled. Such passing of the exception to the outside is also known as bubbling up or propagation.
This feature is more useful when we write complicated code that involves nested structures (e.g., a function calling another function, which may call another function). With the exception re-raising, we can decide where to handle particular exceptions. Certainly, the exact location of handling a specific exception is determined on a case-by-case basis. Here, I can show you how we can re-raise an exception. Let?s see some code first:
Re-raise Exception
In the above code, we have two functions, with run_cast_number calling the other function cast_number. We call the function with a string twice, both of which result in an exception, such that the message ?Failed to cast? is printed because the exception is handled in the cast_number function. However, for the second time, we call the function, we ask the cast_number function to re-raise the exception (Lines 8?9) such that the except clause runs in the run_cast_number function (Lines 15 & 22?23).
11. User-defined exception
In many cases, we can use the built-in exceptions to help us raise and handle exceptions in our project. However, Python gives us the flexibility of creating our own custom exception class. If you don?t know how to create a Python custom class, refer to my previous article on this:
Declare Your First Python Class ? Understand 3 Basic Components
Better organize your code using custom classes
medium.com
Specifically, we need to declare a class as a subclass of the built-in Exception class. Conventionally, you should name your class as something ending with Error (e.g., MediumDataError).
Custom Exception Class
As shown above, we create a custom exception class called FileExtensionError. When we raise such an exception, using the class name alone won?t work, as shown in Lines 10?13. Instead, we should instantiate this exception by setting the two positional arguments for the constructor method. As shown in Line 17, we can see the custom exception message, by implementing the __str__ method. In other words, the exception message is generated by calling the str() function.
12. When to Raise
We?ve learned how to raise built-in and custom exceptions. When we learn Python, most of the time, we only need to know how to handle exceptions. However, with the advancement of your Python skills, you may be wondering when you should raise an exception.
The rule of thumb is you should raise an exception when your code will possibly run into some scenarios when execution can?t proceed. By raising a proper exception, it will allow other parts of your code to handle the exception properly, such that the execution can proceed.
When to Raise Exception
In the above code, we first define a function, read_data, that can read a file. Suppose that the other function process_data is a public API and we don?t have good control over what file type the user is going to pass. Therefore, when we read the data using the read_data function, we want to raise an exception, because our program can?t proceed without the correct data.
We call the public API process_data function twice, with one using the wrong data type and the other using the correct data type. For the former condition, the exception is properly raised and handled such that our program doesn?t crash and the user is also informed of the mistake about the API use.
Conclusions
In this article, we reviewed various aspects regarding the handling and raising of exceptions in Python. By taking exceptions into your project, your code will become more robust and you will be less likely to run into scenarios where execution can?t be recovered.
