Refactoring PowerShell – Switch Statements

Regardless of your experience within technology, the process of creating code usually starts out with a solution that is just enough to get the job done. The solution is then typically tweaked and improved continuously until it is either “production worthy” or “good enough for the requirements“. This process of improvement is called refactoring. Refactoring is a skill that all technical professionals should become proficient in, regardless if you are an IT Pro, a developer or just someone who needs to automate things.

There are many reasons to refactor code, including:

  • Add new features
  • Remove unused features
  • Improve performance
  • Increase readability, maintainability or test-ability
  • Improve design
  • Improve adherence to standards or best practices

Refactoring in Code Reviews

One of the jobs of a code reviewer is to suggest areas that could be refactored to improve some of the areas above. I’ve often found that I’ll suggest the same set of refactorings in my reviews. So rather than just putting the suggesting into the code review, I thought I’d start writing them down here in a series that I could refer contributors to as well as help anyone else who happens to come across these.

Unit Tests and Refactoring

Because refactoring requires changing code, how can we be sure that we’re not breaking functionality or introducing bugs? That is where unit testing comes in. With PowerShell, we’ll typically use the PowerShell Pester module to create unit tests that allow us to more safely refactor code. Unit testing is beyond the scope of this post.

Switch Statement Refactoring

One of the common patterns I’ve come across is PowerShell switch statements being used is to map from one values to another set of values. For example:

This converts a colour name (e.g. red, green, blue, white) to a colour value. If a colour name can not be matched then it returns 0x0 (black). Admittedly, this is a bit of an unlikely example, but it demonstrates the pattern.

This is by no means incorrect or “bad code”. It is completely valid and solves the problem perfectly well. But as you can see this requires a lot of code to perform a simple mapping.

Mapping using a Hash Table

An alternative to using a switch statement is to use a hash table:

This can simplify the code slightly by removing the break statement and braces.

Note: The break statement is not strictly required in this example from a functional perspective, but including them increases overall performance of the switch statement.

You may have noticed that the hash table above does not quite match the behavior of the switch statement: the default mapping to 0x0 is not handled. So, in this case, we’d need to include additional code to handle this:

To handle the default we’re using a quasi null coalescing operator. PowerShell doesn’t have a native null coalescing operator like many languages, but it does have a way of simulating it using the line:

You could definitely argue that using a hash table mapping with a null coalescing operator does not make the code easier to read or maintain. But the purpose here is not to define which approach is best, rather to offer alternative patterns.

One other benefit of using a hash table for mapping is that it can be separated out into a separate psd1 file. This allows editing of the mapping table elements without having to edit the code itself:

The psd1 file containing the mapping data (mapping.psd1):

Reversing the Mapping

How do we use a similar pattern to reverse the mapping? For example, mapping a colour value back to a colour name:

To implement the same functionality using a hash table also including the null coalescing operator you could use:

Using a Hash Table with Script Values

Switch blocks may contain more than just a single statement. For example:

If your switch blocks do more than just perform a mapping, you can assign script blocks to the hash table values instead:

Instead of containing a value in each hash table item, a script block is specified. The Invoke() method can then be called on the script block.

Enumerated Types

If you’re using PowerShell 5.0 or above (you hopefully are), you’re also able to use the enum keyword to define an enumerated type that can also be used to replace switch statements in some situations.

The enumerated type only needs to be declared once.

But what do we need to do if we want to have a default returned if the value is invalid in the mapping? In that case we need to use the TryParse method of the enumerated type to try and parse the value and return a default value if it is invalid:

However, we can’t assign scriptblocks to the values in an enumerated type – only constant values may be assigned. This means we can’t implement scenarios where we’d like to have the value contain more than one instruction. But this shouldn’t be too much of a problem, because if you do find yourself being limited by this, then you should probably be looking to use more advanced object oriented programming patterns such as polymorphism – which is well beyond the scope of this post. But if you’re interested to know more, review this article (not PowerShell specific).

Wrapping Up

This post is all about looking at different ways of writing the same code. It isn’t trying to say which way is better or worse. If you have a preference and it works for you, by all means, keep on using it. This is simply to provide alternative methods that may or may not make code more readable and maintainable.

Feel free to comment with alternative methods of refactoring switch statements.

 

 

 

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s