Why Test-Path doesn't work in HKEY_LOCAL_MACHINE, but only HKLM:

I found behavior in Powershell that looks very inconsistent and confusing, mostly different notation when accessing registry keys and what syntax is expected (as an argument) or supplied (as a return value) by different commands, especially this means:

HKEY_LOCAL_MACHINE ... versus HKLM \

Let me show you an example:

$baseDir = "HKLM:\System\CurrentControlSet\Enum\SCSI"
$Results = Get-ChildItem $baseDir -Recurse -ErrorAction SilentlyContinue

foreach ($item in $Results)
{    
    $Subkey = $item.Name
    $keyExists = Test-Path "$Subkey" -PathType Container -ErrorAction SilentlyContinue
    if ($keyExists -eq $False) 
    {
        New-Item $Subkey            
    }
}

      

        
        
        
      

    

So what happens is this:

$Subkey = $item.Name

      

        
        
        
      

    

returns HKEY_LOCAL_MACHINE \ System \ CurrentControlSet \ Enum \ SCSI \ SomePath

and

$keyExists = Test-Path "$Subkey" -PathType Container -ErrorAction SilentlyContinue

      

        
        
        
      

    

doesn't work with this syntax, i.e. returns "$ false" even though the path exists.

As a workaround, I introduced a sequence of lines of code between these two lines, which fixes the problem:

$Subkey = $Subkey -replace "HKEY_LOCAL_MACHINE", "HKLM:"

      

        
        
        
      

    

This works - it changes the line: HKLM: \ System \ CurrentControlSet \ Enum \ SCSI \ SomePath so Test-Path can work with this syntax, but it's not very smart.

What am I missing at all? Why doesn't powershell return the result names from Get-ChildItem in a way that is suitable for further processing in powershell? Why not always use the same syntax style?

For me this is a design flaw in Powershell, or is there another way to fix this?

(Note: this is just a stripped down example showing the main problem, I know there is no point in looking for children and checking for its existence ...)