Everything you wanted to know about $null

The PowerShell $null often appears to be simple but it has a lot of nuances. Let's take a close look at $null so you know what happens when you unexpectedly run into a $null value.

Note

The original version of this article appeared on the blog written by @KevinMarquette. The PowerShell team thanks Kevin for sharing this content with us. Please check out his blog at PowerShellExplained.com.

What is NULL?

You can think of NULL as an unknown or empty value. A variable is NULL until you assign a value or an object to it. This can be important because there are some commands that require a value and generate errors if the value is NULL.

PowerShell $null

$null is an automatic variable in PowerShell used to represent NULL. You can assign it to variables, use it in comparisons and use it as a place holder for NULL in a collection.

PowerShell treats $null as an object with a value of NULL. This is different than what you may expect if you come from another language.

Examples of $null

Anytime you try to use a variable that you have not initialized, the value is $null. This is one of the most common ways that $null values sneak into your code.

PS> $null -eq $undefinedVariable
True

If you happen to mistype a variable name then PowerShell sees it as a different variable and the value is $null.

The other way you find $null values is when they come from other commands that don't give you any results.

PS> function Get-Nothing {}
PS> $value = Get-Nothing
PS> $null -eq $value
True

Impact of $null

$null values impact your code differently depending on where they show up.

In strings

If you use $null in a string, then it's a blank value (or empty string).

PS> $value = $null
PS> Write-Output "The value is $value"
The value is

This is one of the reasons that I like to place brackets around variables when using them in log messages. It's even more important to identify the edges of your variable values when the value is at the end of the string.

PS> $value = $null
PS> Write-Output "The value is [$value]"
The value is []

This makes empty strings and $null values easy to spot.

In numeric equation

When a $null value is used in a numeric equation then your results are invalid if they don't give an error. Sometimes the $null evaluates to 0 and other times it makes the whole result $null. Here is an example with multiplication that gives 0 or $null depending on the order of the values.

PS> $null * 5
PS> $null -eq ( $null * 5 )
True

PS> 5 * $null
0
PS> $null -eq ( 5 * $null )
False

In place of a collection

A collection allows you use an index to access values. If you try to index into a collection that is actually null, you get this error: Cannot index into a null array.

PS> $value = $null
PS> $value[10]
Cannot index into a null array.
At line:1 char:1
+ $value[10]
+ ~~~~~~~~~~
    + CategoryInfo          : InvalidOperation: (:) [], RuntimeException
    + FullyQualifiedErrorId : NullArray

If you have a collection but try to access an element that is not in the collection, you get a $null result.

$array = @( 'one','two','three' )
$null -eq $array[100]
True

In place of an object

If you try to access a property or sub property of an object that doesn't have the specified property, you get a $null value like you would for an undefined variable. It doesn't matter if the variable is $null or an actual object in this case.

PS> $null -eq $undefined.some.fake.property
True

PS> $date = Get-Date
PS> $null -eq $date.some.fake.property
True

Method on a null-valued expression

Calling a method on a $null object throws a RuntimeException.

PS> $value = $null
PS> $value.toString()
You cannot call a method on a null-valued expression.
At line:1 char:1
+ $value.tostring()
+ ~~~~~~~~~~~~~~~~~
    + CategoryInfo          : InvalidOperation: (:) [], RuntimeException
    + FullyQualifiedErrorId : InvokeMethodOnNull

Whenever I see the phrase You cannot call a method on a null-valued expression then the first thing I look for are places where I am calling a method on a variable without first checking it for $null.

Checking for $null

You may have noticed that I always place the $null on the left when checking for $null in my examples. This is intentional and accepted as a PowerShell best practice. There are some scenarios where placing it on the right doesn't give you the expected result.

Look at this next example and try to predict the results:

if ( $value -eq $null )
{
    'The array is $null'
}
if ( $value -ne $null )
{
    'The array is not $null'
}

If I do not define $value, the first one evaluates to $true and our message is The array is $null. The trap here is that it's possible to create a $value that allows both of them to be $false

$value = @( $null )

In this case, the $value is an array that contains a $null. The -eq checks every value in the array and returns the $null that is matched. This evaluates to $false. The -ne returns everything that doesn't match $null and in this case there are no results (This also evaluates to $false). Neither one is $true even though it looks like one of them should be.

Not only can we create a value that makes both of them evaluate to $false, it's possible to create a value where they both evaluate to $true. Mathias Jessen (@IISResetMe) has a good post that dives into that scenario.

PSScriptAnalyzer and VSCode

The PSScriptAnalyzer module has a rule that checks for this issue called PSPossibleIncorrectComparisonWithNull.

PS> Invoke-ScriptAnalyzer ./myscript.ps1

RuleName                              Message
--------                              -------
PSPossibleIncorrectComparisonWithNull $null should be on the left side of equality comparisons.

Because VS Code uses the PSScriptAnalyser rules too, it also highlights or identifies this as a problem in your script.

Simple if check

A common way that people check for a non-$null value is to use a simple if() statement without the comparison.

if ( $value )
{
    Do-Something
}

If the value is $null, this evaluates to $false. This is easy to read, but be careful that it's looking for exactly what you're expecting it to look for. I read that line of code as:

If $value has a value.

But that's not the whole story. That line is actually saying:

If $value is not $null or 0 or $false or an empty string or an empty array.

Here is a more complete sample of that statement.

if ( $null -ne $value -and
        $value -ne 0 -and
        $value -ne '' -and
        ($value -isnot [array] -or $value.Length -ne 0) -and
        $value -ne $false )
{
    Do-Something
}

It's perfectly OK to use a basic if check as long as you remember those other values count as $false and not just that a variable has a value.

I ran into this issue when refactoring some code a few days ago. It had a basic property check like this.

if ( $object.property )
{
    $object.property = $value
}

I wanted to assign a value to the object property only if it existed. In most cases, the original object had a value that would evaluate to $true in the if statement. But I ran into an issue where the value was occasionally not getting set. I debugged the code and found that the object had the property but it was a blank string value. This prevented it from ever getting updated with the previous logic. So I added a proper $null check and everything worked.

if ( $null -ne $object.property )
{
    $object.property = $value
}

It's little bugs like these that are hard to spot and make me aggressively check values for $null.

$null.Count

If you try to access a property on a $null value, that the property is also $null. The count property is the exception to this rule.

PS> $value = $null
PS> $value.count
0

When you have a $null value, then the count is 0. This special property is added by PowerShell.

[PSCustomObject] Count

Almost all objects in PowerShell have that count property. One important exception is the [PSCustomObject] in Windows PowerShell 5.1 (This is fixed in PowerShell 6.0). It doesn't have a count property so you get a $null value if you try to use it. I call this out here so that you don't try to use .Count instead of a $null check.

Running this example on Windows PowerShell 5.1 and PowerShell 6.0 gives you different results.

$value = [PSCustomObject]@{Name='MyObject'}
if ( $value.count -eq 1 )
{
    "We have a value"
}

Enumerable null

There is one special type of $null that acts differently than the others. I am going to call it the enumerable null but it's really a System.Management.Automation.Internal.AutomationNull. This enumerable null is the one you get as the result of a function or script block that returns nothing (a void result).

PS> function Get-Nothing {}
PS> $nothing = Get-Nothing
PS> $null -eq $nothing
True

If you compare it with $null, you get a $null value. When used in an evaluation where a value is required, the value is always $null. But if you place it inside an array, it's treated the same as an empty array.

PS> $containempty = @( @() )
PS> $containnothing = @($nothing)
PS> $containnull = @($null)

PS> $containempty.count
0
PS> $containnothing.count
0
PS> $containnull.count
1

You can have an array that contains one $null value and its count is 1. But if you place an empty array inside an array then it's not counted as an item. The count is 0.

If you treat the enumerable null like a collection, then it's empty.

If you pass in an enumerable null to a function parameter that isn't strongly typed, PowerShell coerces the enumerable null into a $null value by default. This means inside the function, the value is treated as $null instead of the System.Management.Automation.Internal.AutomationNull type.

Pipeline

The primary place you see the difference is when using the pipeline. You can pipe a $null value but not an enumerable null value.

PS> $null | ForEach-Object{ Write-Output 'NULL Value' }
'NULL Value'
PS> $nothing | ForEach-Object{ Write-Output 'No Value' }

Depending on your code, you should account for the $null in your logic.

Either check for $null first

  • Filter out null on the pipeline (... | Where {$null -ne $_} | ...)
  • Handle it in the pipeline function

foreach

One of my favorite features of foreach is that it doesn't enumerate over a $null collection.

foreach ( $node in $null )
{
    #skipped
}

This saves me from having to $null check the collection before I enumerate it. If you have a collection of $null values, the $node can still be $null.

The foreach started working this way with PowerShell 3.0. If you happen to be on an older version, then this is not the case. This is one of the important changes to be aware of when back-porting code for 2.0 compatibility.

Value types

Technically, only reference types can be $null. But PowerShell is very generous and allows for variables to be any type. If you decide to strongly type a value type, it cannot be $null. PowerShell converts $null to a default value for many types.

PS> [int]$number = $null
PS> $number
0

PS> [bool]$boolean = $null
PS> $boolean
False

PS> [string]$string = $null
PS> $string -eq ''
True

There are some types that do not have a valid conversion from $null. These types generate a Cannot convert null to type error.

PS> [datetime]$date = $null
Cannot convert null to type "System.DateTime".
At line:1 char:1
+ [datetime]$date = $null
+ ~~~~~~~~~~~~~~~~~~~~~~~
    + CategoryInfo          : MetadataError: (:) [], ArgumentTransformationMetadataException
    + FullyQualifiedErrorId : RuntimeException

Function parameters

Using a strongly typed values in function parameters is very common. We generally learn to define the types of our parameters even if we tend not to define the types of other variables in our scripts. You may already have some strongly typed variables in your functions and not even realize it.

function Do-Something
{
    param(
        [String] $Value
    )
}

As soon as you set the type of the parameter as a string, the value can never be $null. It's common to check if a value is $null to see if the user provided a value or not.

if ( $null -ne $Value ){...}

$Value is an empty string '' when no value is provided. Use the automatic variable $PSBoundParameters.Value instead.

if ( $null -ne $PSBoundParameters.Value ){...}

$PSBoundParameters only contains the parameters that were specified when the function was called. You can also use the ContainsKey method to check for the property.

if ( $PSBoundParameters.ContainsKey('Value') ){...}

IsNotNullOrEmpty

If the value is a string, you can use a static string function to check if the value is $null or an empty string at the same time.

if ( -not [string]::IsNullOrEmpty( $value ) ){...}

I find myself using this often when I know the value type should be a string.

When I $null check

I am a defensive scripter. Anytime I call a function and assign it to a variable, I check it for $null.

$userList = Get-ADUser kevmar
if ($null -ne $userList){...}

I much prefer using if or foreach over using try/catch. Don't get me wrong, I still use try/catch a lot. But if I can test for an error condition or an empty set of results, I can allow my exception handling be for true exceptions.

I also tend to check for $null before I index into a value or call methods on an object. These two actions fail for a $null object so I find it important to validate them first. I already covered those scenarios earlier in this post.

No results scenario

It's important to know that different functions and commands handle the no results scenario differently. Many PowerShell commands return the enumerable null and an error in the error stream. But others throw exceptions or give you a status object. It's still up to you to know how the commands you use deal with the no results and error scenarios.

Initializing to $null

One habit that I have picked up is initializing all my variables before I use them. You are required to do this in other languages. At the top of my function or as I enter a foreach loop, I define all the values that I'm using.

Here is a scenario that I want you to take a close look at. It's an example of a bug I had to chase down before.

function Do-Something
{
    foreach ( $node in 1..6 )
    {
        try
        {
            $result = Get-Something -ID $node
        }
        catch
        {
            Write-Verbose "[$result] not valid"
        }

        if ( $null -ne $result )
        {
            Update-Something $result
        }
    }
}

The expectation here is that Get-Something returns either a result or an enumerable null. If there's an error, we log it. Then we check to make sure we got a valid result before processing it.

The bug hiding in this code is when Get-Something throws an exception and doesn't assign a value to $result. It fails before the assignment so we don't even assign $null to the $result variable. $result still contains the previous valid $result from other iterations. Update-Something to execute multiple times on the same object in this example.

I set $result to $null right inside the foreach loop before I use it to mitigate this issue.

foreach ( $node in 1..6 )
{
    $result = $null
    try
    {
        ...

Scope issues

This also helps mitigate scoping issues. In that example, we assign values to $result over and over in a loop. But because PowerShell allows variable values from outside the function to bleed into the scope of the current function, initializing them inside your function mitigates bugs that can be introduced that way.

An uninitialized variable in your function is not $null if it's set to a value in a parent scope. The parent scope could be another function that calls your function and uses the same variable names.

If I take that same Do-something example and remove the loop, I would end up with something that looks like this example:

function Invoke-Something
{
    $result = 'ParentScope'
    Do-Something
}

function Do-Something
{
    try
    {
        $result = Get-Something -ID $node
    }
    catch
    {
        Write-Verbose "[$result] not valid"
    }

    if ( $null -ne $result )
    {
        Update-Something $result
    }
}

If the call to Get-Something throws an exception, then my $null check finds the $result from Invoke-Something. Initializing the value inside your function mitigates this issue.

Naming variables is hard and it's common for an author to use the same variable names in multiple functions. I know I use $node,$result,$data all the time. So it would be very easy for values from different scopes to show up in places where they should not be.

Redirect output to $null

I have been talking about $null values for this entire article but the topic is not complete if I didn't mention redirecting output to $null. There are times when you have commands that output information or objects that you want to suppress. Redirecting output to $null does that.

Out-Null

The Out-Null command is the built-in way to redirect pipeline data to $null.

New-Item -Type Directory -Path $path | Out-Null

Assign to $null

You can assign the results of a command to $null for the same effect as using Out-Null.

$null = New-Item -Type Directory -Path $path

Because $null is a constant value, you can never overwrite it. I don't like the way it looks in my code but it often performs faster than Out-Null.

Redirect to $null

You can also use the redirection operator to send output to $null.

New-Item -Type Directory -Path $path > $null

If you're dealing with command-line executables that output on the different streams. You can redirect all output streams to $null like this:

git status *> $null

Summary

I covered a lot of ground on this one and I know this article is more fragmented than most of my deep dives. That is because $null values can pop up in many different places in PowerShell and all the nuances are specific to where you find it. I hope you walk away from this with a better understanding of $null and an awareness of the more obscure scenarios you may run into.