Tag Archives: Select-Object

A Variable’s Current and Previous Value

How to Store a Variable’s Previous Value with the New Value

There was a recent post on the PowerShell Facebook group that said this: “Is there a function or option to check previous $Global:variable -ne Current $Global:variable?” The person that created the post wanted to know if there’s a way to recall the last value a variable stored, once it had already been assigned something new. Here’s my answer, and potential solution.

You have to think about a variable’s previous assignment, as if it never existed, even though you know that variable held something, or was assigned something, previously. Once its value is gone, it’s gone. Now, all that said, I came up with an option. The option, is to use a variable’s description property to store its previous value.

Let’s begin by checking the value of an uninitialized variable, and then assign it something.

PS > $x
PS > # No value.
PS > $x = 'first value'
PS > $x
first value

Now that our variable has a value, let’s take a look at all the variable’s properties using the Get-Variable cmdlet. Notice the Description property; we’re going to use this in a less than conventional way.

PS > Get-Variable -Name x | Select-Object -Property *

Name        : x
Description :
Value       : first value
Visibility  : Public
Module      :
ModuleName  :
Options     : None
Attributes  : {}

Using the Set-Variable cmdlet, we’ll make two modifications to our variable, at nearly the same time. First, we’ll update the Description property, so that it’s holding the original variable assignment (the string ‘first value’). In the same, Set-Variable command, we’ll modify the current value of the variable to the string ‘second value.’ Notice in the final below command, that our Description property has a value now, too.

PS > Set-Variable -Name x -Value 'second value' -Description $x
PS > $x
second value
PS > Get-Variable -Name x | Select-Object -Property *

Name        : x
Description : first value
Value       : second value
Visibility  : Public
Module      :
ModuleName  :
Options     : None
Attributes  : {}

In the next example, you’ll see how we can return the original value and the variable’s updated value. Beneath that, I’ve included a couple ways to compare these values. That gets back to what the person on Facebook was trying to determine. Again, this is fairly unconventional use of the Description property, but it does avoid the need for a second variable to hold the first variable’s original value. That said, one of my examples uses a second, comparison variable.

PS > (Get-Variable -Name x).Description
first value
PS > (Get-Variable -Name x).Value
second value
PS > $x
second value

PS > (Get-Variable -Name x).Description -ne (Get-Variable -Name x).Value
True
PS > (Get-Variable -Name x).Description -ne $x
True
PS > $y = (Get-Variable -Name x).Description
PS > $y -ne $x
True

In these last examples, we’re running into a bit of a problem, we are going to have to keep in mind (if anyone even dares use this approach). When we take a value that isn’t a string, and place it into the Description property, it becomes a string. That means, that when we take it back out, we’ll need to cast it back to its proper type. Take a look at the next series of examples for some assistance with this concept.

PS > $z = 5
PS > Set-Variable -Name z -Value 10 -Description $z
PS > Get-Variable -Name z | Select-Object -Property *

Name        : z
Description : 5
Value       : 10
Visibility  : Public
Module      :
ModuleName  :
Options     : None
Attributes  : {}

PS > (Get-Variable -Name z).Description | Get-Member |
>>> Select-Object -Property TypeName -Unique

TypeName
--------
System.String

PS > [int](Get-Variable -Name z).Description
5
PS > [int](Get-Variable -Name z).Description | Get-Member |
>>> Select-Object -Property TypeName -Unique

TypeName
--------
System.Int32

This goes for decimals values, too. If you use the Description property to store something that’s not a string, then you’re going to have to properly cast it when you’re taking it out, or you’re going to have a string. In these next examples, I used the GetType method to determine the value’s type, as opposed to Get-Member and Select-Object used above.

PS > $m = 2.5
PS > $m.GetType().Name
Double
PS > Set-Variable -Name m -Value 3.5 -Description $m
PS > $m
3.5
PS > $m.GetType().Name
Double
PS > (Get-Variable -Name m).Description
2.5
PS > (Get-Variable -Name m).Description.GetType().Name
String
PS > [double](Get-Variable -Name m).Description
2.5
PS > ([double](Get-Variable -Name m).Description).GetType().Name
Double

And that’s it. It’s may not be the first choice for saving a previous variable’s value, but it’s a choice. I rather liked the array option that was recommended; however, like the $Error array, I’d be tempted to put the newest/current value of a variable into index 0 and not at the end of the array. Anyway, back to real life now.

View Current PowerShell.org Q&A Forum Topics

Note: Update added at the bottom on this post on August, 9, 2016. Please read.

Sometimes you don’t always have the time to finish something you’ve started. For me, it was this function. I pounded this out in a quick few minutes, and while I don’t see myself investing in it any further, I didn’t want to forget the function, and thought I would hang on to it somewhere. Well, that’s why it’s here, especially as someone may find it useful, or helpful.

The function, which I called Get-PowerShell.orgForumTopic, runs out to PowerShell.org and grabs the current topics (page one) from the PowerShell Q&A forum (http://powershell.org/wp/forums/forum/windows-powershell-qa). It only returns the Thread name and the URL, because, well, that’s what seemed useful and relevant at the time I wrote it (which was many months ago).

Function Get-PowerShell.orgForumTopic {
    [CmdletBinding()]
    Param ()

    Begin {
    } # End Begin.

    Process {
        (Invoke-WebRequest -Uri 'http://powershell.org/wp/forums/forum/windows-powershell-qa/' |
            Select-Object -ExpandProperty Links |
            Where-Object {$_.outerHTML -like '*http://powershell.org/wp/forums/topic*'} |
            Select-Object @{N='Thread';E={$_.innerHTML}},@{N='Url';E={$_.href}} |
            Select-Object -First 30)[(0..30 |
                ForEach-Object {
                    If (-not($_ % 2)) {
                        $_
                    }
                 }
            )]
    } # End Process.

    End {
    } # End End.
} # End Function: Get-PowerShell.orgForumTopic

Here’s what the results looked like in the ConsoleHost, near in time to when this post was published.

current-powershell.org-qa-forum-topics-01

While I never added any more to this function, I had some ideas: add the thread status, add the “started by user,” add the user that made the last post, add the number of posts per topic, and allow it to run against other PowerShell.org forum topics. It might’ve also been helpful to include additional pages, if requested by the user of the function, such as adding a -Pages parameter (-Pages 4).

Anyway, here it is. Beside being helpful to see the top PowerShell.org Q&A forum posts at a specific point in time, it’s an interesting example of reading from a webpage, of which I had minimal experience. So yeah, I probably learned something by doing this exercise.

If you want to do the same, then start by running the first command, Invoke-WebRequest -Uri ‘http://powershell.org/wp/forums/forum/windows-powershell-qa/’. Then add the pipe and first Select-Object command and run that. Then add the Where-Object command, and so on. This will allow you to see how I finally got to only returning the “Threads” and “Urls.” Take care.

Update: Since a site redesign at PowerShell.org, this function, no longer functions. I’m not sure that I’ll bother to update it — I don’t get the feeling that it was ever used by anyone — but I’ll keep this post up for anything else it may offer, that may be helpful in learning PowerShell.

Add, Concatenate, and a Little More

My desire to know Windows PowerShell from beginning to end isn’t just for me. It’s also for those around me — online, and in real life. I want to be a go-to-guy for all things PowerShell. I’ve found something I really love about the industry right now, and so I want know it well, and share it often. I truly believe that learning the fundamentals of PowerShell make one much more prepared to use the language interactively, as well as inside scripts and functions. I’ve said it before: It’s easy to spot someone that doesn’t know the fundamentals of PowerShell, and is only out to learn as much as is required to complete their current task.

A couple weeks ago a coworker asked me why their conditional wasn’t working (think, the part inside the parentheses in an If statement, for example). While I never saw the code, as the conversation happen over cube walls and by instant message, I suspected they were trying to compare a numeric value with something that wasn’t numeric. It turned out I was right. I determined this, the moment they said they had quotes around their numeric value.

It’s simple. The moment we put numeric values inside quotes — single or otherwise — we are no longer working numbers — we’re working with strings. Take a look at the example below. In the first command, we use addition to add the numeric values 5 and 10, which results in the (numeric) value of 15. In the second command, we concatenate (think, combine or join) — using the same operator (+). When we concatenate the string value of 5 with the string value of 10, we create the (string) value of 510.

PS> 5 + 10
15
PS> '5' + '10'
510

The information I sent to my colleague by IM is listed below. I thought it would best explain the difference between a “numeric value” inside quotes, and one not inside quotes. In this example, 123 is an integer (System.Int32) without quotes around it, and a string (System.String) with quotes around it. A quick note about Select-Object’s -Unique switch parameter: This parameter returns duplicated results once, and only once. Had this not been included, it would have returned the same result for each member of the object (each method, property, etc.).

PS> 123 | Get-Member | Select-Object TypeName -Unique

TypeName
--------
System.Int32

PS> '123' | Get-Member | Select-Object TypeName -Unique

TypeName
--------
System.String

In our first example in this post, we saw what happens when the add two numeric values and what happens when we concatenate two strings. If you think like me, you’re probably wondering what happens when we concatenate a string value to a numeric value, and add a numeric value to a string character. It depends. Take a look at the slightly, modified example below, and then we’ll discuss it.

PS> 5 + '10'
15
PS> '5' + 10
510

Did you figure it out what’s happened in those examples? In the first command, 5 + ’10’, the numeric on the left of the equation, has forced the string on the right side, to switch from a string to a numeric data type. With two numeric values, it then added the two values and returned 15.

The second command, or equation, works much like the first one, except in reverse. The string character 5 on the left, forces the numeric value on the right into being a string and then concatenates the two, which results in 510 again. Whatever is on the left is going to try and change the type of whatever is on the right — when it can. So, when can’t it?

PS> 5 + '1a2b3c'
Cannot convert value "1a2b3c" to type "System.Int32". Error: "Input string was not in a correct format."
...
PS> '1a2b3c' + 10
1a2b3c10

The first command in the example above fails because it cannot convert the string, ‘1a2b3c,’ into a numeric value because it includes non-numeric characters — makes sense to me. Notice that this error helps prove what I said earlier: The data type of value on the left is being used to try and change the data type of the value on the right. The second equation works, because again, we’re simply joining, or concatenating, two values.

PS> 'Have a great ' + 2016 + '!'
Have a great 2016!

In closing, I’d like to add a word about the concatenation operation just above. While the plus sign, or rather, the concatenation operator does what it does well, it’s not used nearly as much as it was in the past for building strings — at least, I don’t see it much anymore. In fact, when I do see it, I almost always assume that what I’m looking at was written long ago and then borrowed. What you’re more likely to see these days, is in the example below.

PS> $Year = 2016 # It's a numeric value (no quotes).
PS> # Use double-quotes (below) to display the variable's value inside the string.
PS> "Have a great $Year!" # No concatenation operators needed!
Have a great 2016!

Thanks for the reading my first post of 2016!

Prep computers.txt File for HTA

Last week, I opted to share a couple of old HTAs I had written. HTAs are HTML Applications and allow administrations the ability to create a graphical interface for their scripts. It’s an older technology and not something I still write, or use. Even so, I wanted to share them with the community. As they’re not Windows PowerShell-related, I thought I should circle back on that post and incorporate some PowerShell.

The second of the two HTAs I shared was called Remote Desktop Assistant (download link: RDAssistantv2.1 (299 downloads) ). Its purpose is to allow a user to select a computer description from a list and open Remote Desktop to connect to that computer. I know, I know, this goes against all things PowerShell, but it was written a long time ago. The HTA has a requirement for an external text file called computers.txt that stores computer descriptions and computer names / IP addresses, such as we have in the list below. It’s a computer description, a semi-colon, and the computer name or IP address.

computer1;10.10.10.5
computer2;dns1.mydomain.com
computer3;10.10.10.9

We can make use of Active Directory and PowerShell to create this list and subsequent text file, so we don’t have to do it manually. In fact, once you had your code written you could schedule it to ensure the computers.txt file was as accurate as the last time the scheduled task ran. While the preferred way to do this would be to store host names (at least in my opinion), I’ll show examples of collecting the IP addresses, too.

In the first example, we’ll pull all of our servers from a single Organizational Unit to return the names and DNSHostNames.

PS> Get-ADComputer -Filter * -SearchBase 'OU=Servers,DC=mydomain,DC=com' -Properties DNSHostName | Select-Object Name,DNSHostName

Name                                                        DNSHostName
----                                                        -----------
SQL01                                                       SQL01.mydomain.com
SQL02                                                       SQL02.mydomain.com
SQL03                                                       SQL03.mydomain.com
SQL04                                                       SQL04.mydomain.com

In this example, we return the names and IP addresses.

PS> Get-ADComputer -Filter * -SearchBase 'OU=Servers,DC=mydomain,DC=com' -Properties IPv4Address | Select-Object Name,IPv4Address

Name                                                        IPv4Address
----                                                        -----------
SQL01                                                       10.10.10.30
SQL02                                                       10.10.10.31
SQL03                                                       10.10.10.32
SQL04                                                       10.10.10.33

While these are the results we want, we need to get them into the proper format. We’ll do this by looping through each result and concatenating the two properties with semi-colon in between. To do this, we do not need to use the Select-Object cmdlet to return the Name and IPv4Address, or DNSHostName.

PS> Get-ADComputer -Filter * -SearchBase 'OU=Servers,DC=mydomain,DC=com' -Properties IPv4Address | ForEach-Object {"$($_.Name):$($_.IPv4Address)"}
SQL01;10.10.10.30
SQL02;10.10.10.31
SQL03;10.10.10.32
SQL04;10.10.10.33

With the host name set, we’ll take this one step further and create our computers.txt file.

PS> Get-ADComputer -Filter * -SearchBase 'OU=Servers,DC=mydomain,DC=com' -Properties DNSHostName | ForEach-Object {"$($_.Name):$($_.DNSHostName)"} | Out-File -FilePath C:\computers.txt
PS> Get-Content -Path C:\computers.txt
SQL01;SQL01.mydomain.com
SQL02;SQL02.mydomain.com
SQL03;SQL03.mydomain.com
SQL04;SQL04.mydomain.com

Chances are good that if you use the DNSHostName, you’re never going to have an Active Directory computer object returned without one. The same can’t be said if you use the IPv4Address, as this property is created at the time the results are returned (it queries DNS). Think about it, have you ever seen an IPv4Address property inside Active Directory Users and Computer when viewing a computer object? The DNSHostName option might be the better option, but I’ll leave that up to you.

Return File Sizes in Bytes, KBs, MBs, and GBs, at the Same Time

I scanned a recent forum post on PowerShell.org where the user seemed (again, I scanned it) to want to run some files through Get-ChildItem and Select-Object, and report the size in something other than the bytes default. It didn’t appear he, or she, wanted all the files in one of the measurement groups: B, KB, MB, or GB, but instead in their proper measurement group, dependent on the file’s size.

That may be hard to understand at first read, so let me try and explain it another way. If a file has less than 1,024 bytes, then it should be reported in bytes, if it has 1,024 – 1,048,575 bytes, then it should be reported in Kilobytes, if it has 1,048,576 – 1,073,741,824 bytes then it should be reported Megabytes, and if it has 1,073,741,825 or more bytes, then it should be reported in Gigabytes.

I wrote a long, “one-liner” to do this and have included it, and some sample output, below. I don’t profess to guarantee this doesn’t have any errors, so if you find some, then let me know. In addition, it shows a nice example of a switch statement inside the Expression portion of a calculated property. Fancy.

Get-ChildItem -Path 'C:\test' |
    Select-Object Name,
        @{L='Size';E={
                switch ($_.Length) {
                    # Bytes
                    {$_ -eq 0} {"$('{0:N2}' -f $_) bytes"; break}
                    {$_ -eq 1} {"$('{0:N2}' -f $_) byte"; break}
                    {($_ -gt 1) -and ($_ -le 1023)} {"$('{0:N2}' -f $_) bytes"; break}

                    # Kilobytes
                    {$_ -eq 1024} {"$('{0:N2}' -f ($_/1KB)) Kilobyte"; break}
                    {($_ -gt 1024) -and ($_ -le 1048575)} {"$('{0:N2}' -f ($_/1KB)) Kilobytes"; break}

                    # Megabytes
                    {$_ -eq 1048576} {"$('{0:N2}' -f ($_/1MB)) Megabyte"; break}
                    {($_ -gt 1048576) -and ($_ -le 1073741824)} {"$('{0:N2}' -f ($_/1MB)) Megabytes"; break}

                    # Gigabytes
                    {$_ -eq 1073741825} {"$('{0:N2}' -f ($_/1GB)) Gigabyte"; break}
                    {$_ -gt 1073741825} {"$('{0:N2}' -f ($_/1GB)) Gigabytes"; break}

                    default {Write-Warning -Message 'Unknown Error.'}
                }
            }
        } | Format-Table -AutoSize

Name           Size
----           ----
My-A-File.txt  1.00 byte
My-A-File2.txt 445.00 bytes
My-B-File.txt  1.00 Kilobyte
My-B-File2.txt 1.30 Kilobytes
My-C-File.txt  2.09 Megabytes
My-D-File.txt  3.59 Gigabytes

I can easily see how someone might want to remove the singular vs. plural: Kilobyte vs. Kilobytes, and just use B, KB, MB, and GB, and so…

Update: I gave this some more thought today, and I really didn’t feel like it was okay to not include a modified version that uses abbreviations (B, KB, MB, GB) instead of the full words, as in the previous example. It has removed some complexity, as well, as you can see below.

Get-ChildItem -Path 'C:\test' |
    Select-Object Name,
        @{L='Size';E={
                switch ($_.Length) {
                    # Bytes
                    {($_ -ge 0) -and ($_ -le 1023)} {"$('{0:N2}' -f $_) B"; break}

                    # Kilobytes
                    {($_ -ge 1024) -and ($_ -le 1048575)} {"$('{0:N2}' -f ($_/1KB)) KB"; break}

                    # Megabytes
                    {($_ -ge 1048576) -and ($_ -le 1073741824)} {"$('{0:N2}' -f ($_/1MB)) MB"; break}

                    # Gigabytes
                    {$_ -ge 1073741825} {"$('{0:N2}' -f ($_/1GB)) GB"; break}

                    default {Write-Warning -Message 'Unknown Error.'}
                }
            }
        } | Format-Table -AutoSize

Name           Size
----           ----
My-A-File.txt  1.00 B
My-A-File2.txt 445.00 B
My-B-File.txt  1.00 KB
My-B-File2.txt 1.30 KB
My-C-File.txt  2.09 MB
My-D-File.txt  3.59 GB

Give your Variable a Description

When we think of variables in Windows PowerShell, we often only consider two parts: the name of the variable, and the value of the variable. Well, it turns out that those aren’t the only properties; let’s discuss the description property.

I can’t say I’ve ever added a description to a variable until now, but I can think of times, when looking at someone’s code, where I wish I knew the reason why a variable existed. Best practice would indicate we use meaningful names for our variables, so it’s probably safe to assume that the same people that might use a poorly-named variable, probably don’t know or care that variables can have descriptions… but, I digress.

In this first example, we’ll create a new variable using what I’d call the standard method — the equal sign assignment operator.

PS C:\> $a = 12345
PS C:\> $a
12345
PS C:\> Get-Variable a | Select-Object *

Name        : a
Description :
Value       : 12345
Visibility  : Public
Module      :
ModuleName  :
Options     : None
Attributes  : {}

In line 1, we created a new variable, $a, and set its value to the numeric 12345. In the next line, we simply echoed the variable’s value. In line 4, we used the Get-Variable cmdlet and piped it to the Select-Object cmdlet and * so we were able to view all of its properties. If you take a look at the results, you’ll see the description property, and notice that it’s currently blank.

In the next example, we’ll use the Set-Variable cmdlet to modify the variable’s description.

PS C:\> Set-Variable a -Description 'This variable contains a 5-digit number.'
PS C:\> Get-Variable a | Select-Object *

Name        : a
Description : This variable contains a 5-digit number.
Value       : 12345
Visibility  : Public
Module      :
ModuleName  :
Options     : None
Attributes  : {}

PS C:\>

Once we use the Get-Variable cmdlet, and return all the properties again, we can see that our variable now contains a description value. Now, while many user-defined variables, such as we created here, don’t often contain descriptions, many of the automatic and preference variables do — take a look at a segment of the results returned by the next command.

PS C:\> Get-ChildItem variable: | Select-Object Name,Description | Format-Table -AutoSize

Name                       Description
----                       -----------
...
MaximumAliasCount          Maximum number of aliases allowed in a session
MaximumDriveCount          Maximum number of drives allowed in a session
MaximumErrorCount          Maximum number of errors to retain in a session
MaximumFunctionCount       Maximum number of functions allowed in a session
MaximumHistoryCount        Maximum number of history objects to retain in a session
MaximumVariableCount       Maximum number of variables allowed in a session
MyInvocation
NestedPromptLevel          Dictates what type of prompt should be displayed for the current nesting level
...

Well, there you go: something you might not have known about, and something you may never use.

Get Some Help for Later Reading

I’m just starting to get my hands wet with Microsoft Lync. As I often do, I use the Windows PowerShell cmdlets to help learn more about a product; I did this same thing with Hyper-V. The unfortunate thing about the GUI (think, an MMC snap-in for instance), is that the menu options don’t always tell you exactly what that option is going to do, or its exact purpose. While the GUI can be unclear, PowerShell tells you exactly what a cmdlet does. With that knowledge, I’ve often been able to relate a cmdlet, and it’s purpose, to its respective menu option in the GUI.

I wanted to read though the Lync cmdlet’s help files, but only the cmdlet name and its synopsis. Here’s the command I ran to extract this information for later reading.

PS C:> Get-Command -Module Lync | Get-Help | Select-Object Name,Synopsis | Export-Csv -Path C:\LyncCmdlets.csv -NoTypeInfomation

The command above works this way: it returns all the cmdlets included in the Lync Module and sends (pipes) those to the Get-Help cmdlet. The Get-Help cmdlet pipes it’s results to Select-Object which filters the returned properties to just the Name and Synopsis from each cmdlets’ help file. At the end, those filtered results are sent to Export-Csv which creates a file I can read at my leisure.

Although there’s over 500 cmdlets, its safe to say that many of the nouns (the part after the dash (-)) will be the same across some of the cmdlets. That means that many of the Get-* cmdlets will have a partnering Set-* cmdlet. Get reads information and Set changes it. These nouns may also have a matching New-*, Remove-*, and possibly even a Test-* cmdlet. Now off to do some reading…

What PowerShell Modules does the RSAT Provide (on Windows 8.1)

I built a new Windows 8.1 machine recently, and as we all know, doing that requires new software installations. In my case, that included the RSAT. The RSAT, or Remote Server Administration Tools, allow IT admins the ability to remotely manage features on Windows Server operating systems from a client operating system—here’s a link to the RSAT for Windows 8.1.

Prior to running the RSAT installer, I wanted to collect the currently available Windows PowerShell modules that I already had on my computer. This would allow me to know exactly what modules were added after the RSAT installer finished. The command below collects all the modules, returns only their name, and then drops that into a file on my computer. We’ll use the file in a moment, as part of a comparison.

PS C:\> Get-Module -ListAvailable | Select-Object -ExpandProperty Name | Out-File C:\Users\tommymaynard\Pre-RSAT-PowerShell-Modules.txt

I verified the file contained the module names by using the Get-Content cmdlet. In addition, I ran a slightly modified command to get the item count in the file—both can be seen below. The output created by the first command has been truncated to save space.

PS C:\> Get-Content C:\Users\tommymaynard\Pre-RSAT-PowerShell-Modules.txt
AppBackgroundTask
AppLocker
Appx
AssignedAccess
...
PS C:\> (Get-Content C:\Users\tommymaynard\Pre-RSAT-PowerShell-Modules.txt).Count
57

Once this file was in placed, and I was satisfied that it contained what I wanted, I went ahead with the RSAT install. Upon completion, I reran the Get-Module command above, after modifying the name of the file it would create (pre vs. post). I then read in the contents of the new file (which has been truncated again), and checked the number of modules listed in the file. There were now 75 modules where there had only been 57 before the RSAT install.

PS C:\> Get-Module -ListAvailable | Select-Object -ExpandProperty Name | Out-File C:\Users\tommymaynard\Post-RSAT-PowerShell-Modules.txt
PS C:\> Get-Content C:\Users\tommymaynard\Post-RSAT-PowerShell-Modules.txt
ActiveDirectory
AppBackgroundTask
AppLocker
Appx
...
PS C:\> (Get-Content C:\Users\tommymaynard\Post-RSAT-PowerShell-Modules.txt).Count
75

Knowing that there’s 18 new modules is helpful, but which ones were added? The Compare-Object cmdlet can tell us. As seen below, we have the cmdlet read in the lines of each file and then determine which ones aren’t included in both files.

PS C:\> Compare-Object -ReferenceObject (Get-Content C:\Users\tommymaynard\Desktop\Pre-RSAT-PowerShell-Modules.txt) -DifferenceObject (Get-Content C:\Users\tommymaynard\Desktop\Post-RSAT-PowerShell-Modules.txt)

InputObject                                                 SideIndicator
-----------                                                 -------------
ActiveDirectory                                             =>
BestPractices                                               =>
ClusterAwareUpdating                                        =>
DFSN                                                        =>
DFSR                                                        =>
DhcpServer                                                  =>
DnsServer                                                   =>
FailoverClusters                                            =>
GroupPolicy                                                 =>
IpamServer                                                  =>
IscsiTarget                                                 =>
NetworkLoadBalancingClusters                                =>
NFS                                                         =>
RemoteAccess                                                =>
RemoteDesktop                                               =>
ServerManager                                               =>
ServerManagerTasks                                          =>
UpdateServices                                              =>

As we can see above, our Compare-Object results indicate that the difference object (the post file, or the file on the right) has new modules—now we know which ones were added since installing the RSAT.

Write-Host, Does it Have a Place?

I’ve been working on an advanced function that I can’t wait to share (and no, it’s not the one in this post). I really think it’s something that the Windows PowerShell community has been missing. Okay fine, maybe it’s just something I’ve been missing.

I noticed in development that my Write-Output messages to the user were crossing the pipeline, when the custom object (created by the function), was passed to Select-Object (in certain manners). I have a function below that does the same thing as the one in development.

Here’s how this thing works: The function requires the user to provide the value Write-Output (or, wo), or Write-Host (or, wh) for the -Option parameter. This will determine how the message to the user is written inside the Begin block. The only other thing that’s happening in this function, is that a custom object is being created in the Process block, based on some properties of Win32_BIOS.

Function Test-TMWriteOutputVsHost {
    [CmdletBinding()]
    Param (
    [Parameter(Mandatory=$true,
        HelpMessage="Enter Write-Output or Write-Host")]
    [ValidateSet('Write-Output','wo','Write-Host','wh')]
    [string]$Option
    )

    Begin {
        If ($Option -eq 'Write-Output' -or $Option -eq 'wo') {
            Write-Output 'Inside the Begin block (using Write-Output).'

        } ElseIf ($Option -eq 'Write-Host' -or $Option -eq 'wh') {
            Write-Host 'Inside the Begin block (using Write-Host).'
        }
    } # End Begin

    Process {
        $CollectionVariable = Get-WmiObject -Class Win32_BIOS
        $Object = @()
        $Object += [pscustomobject]@{
            Manufacturer = $CollectionVariable.Manufacturer;
            Name = $CollectionVariable.Name;
            Version = $CollectionVariable.Version
        }
    } # End Process

    End {
        Write-Output $Object
    }
    # End, End
} # End Function

As we can see below, everything works great with both Write-Output or Write-Host, when we don’t pipe the function to the Select-Object cmdlet.

PS C:\> Test-TMWriteOutputVsHost -Option Write-Output
Inside the Begin block (using Write-Output).

Manufacturer                            Name                                    Version
------------                            ----                                    -------
Dell Inc.                               BIOS Date: 08/27/13 11:12:44 Ver: A1... DELL   - 1072009

PS C:\> Test-TMWriteOutputVsHost -Option Write-Host
Inside the Begin block (using Write-Host).

Manufacturer                            Name                                    Version
------------                            ----                                    -------
Dell Inc.                               BIOS Date: 08/27/13 11:12:44 Ver: A1... DELL   - 1072009

Now, let’s pipe our object to some variations of the Select-Object cmdlet and watch some things blow up (when using Write-Output).

PS C:\> # Use the horizontal scrollbar to see the results...
PS C:\> Test-TMWriteOutputVsHost -Option Write-Output | select *

                                                                                                                 Length
                                                                                                                 ------
                                                                                                                     44

PS C:\> Test-TMWriteOutputVsHost -Option Write-Output | select Name,Ver*

Name                                                        Ver*
----                                                        ----

BIOS Date: 08/27/13 11:12:44 Ver: A13.00

PS C:\> Test-TMWriteOutputVsHost -Option Write-Host | select *
Inside the Begin block (using Write-Host).

Manufacturer                            Name                                    Version
------------                            ----                                    -------
Dell Inc.                               BIOS Date: 08/27/13 11:12:44 Ver: A1... DELL   - 1072009

PS C:\> Test-TMWriteOutputVsHost -Option Write-Host | select Name,Ver*
Inside the Begin block (using Write-Host).

Name                                                        Version
----                                                        -------
BIOS Date: 08/27/13 11:12:44 Ver: A13.00                    DELL   - 1072009

The problem with Write-Output, is that the object it’s producing is crossing our pipeline and causing unpredictable behavior—something we can’t include in a function we want to distribute. Here’s the proof: When we pipe our function to Get-Member, we reveal the objects that show up on the other side, and we don’t want the string object coming over with us.

PS C:\> Test-TMWriteOutputVsHost -Option Write-Output | Get-Member | Select-Object TypeName -Unique

TypeName
--------
System.String
System.Management.Automation.PSCustomObject

PS C:\> Test-TMWriteOutputVsHost -Option Write-Host | Get-Member | Select-Object TypeName -Unique
Inside the Begin block (using Write-Host).

TypeName
--------
System.Management.Automation.PSCustomObject

I don’t profess to know it all, so if there’s a way to get around this using Write-Output, then I’d love to hear about it. While I haven’t tried it, I suspect I may be able to create an embedded function—that might be the trick I need. Perhaps I’ll play with that option at another time. Thanks for reading!

Oh, and before someone mentions it, I explained how I feel about Write-Verbose in the comments on a post by Adam Bertram: http://www.adamtheautomator.com/use-write-host-lot.

Using Replace() to Fix Split() (and Convert-Path)

I was working on a recent project that required scripting ACLs, and so I had a Windows PowerShell console open in addition to the ISE. This allowed me to quickly check the owner of a directory (or folder). I could press the up arrow to rerun my command, and I could quickly see if the owner had changed according to my script. Let’s say my directory is called ‘TestFolder’ and is located at the root of the C:\ drive.

PS C:\> New-Item -Path TestFolder -ItemType Directory

    Directory: C:\

Mode                LastWriteTime     Length Name
----                -------------     ------ ----
d----        10/21/2014   9:19 PM            TestFolder

To get the ACL information for the folder, you can use the Get-Acl cmdlet, as in the example below.

PS C:\> Get-Acl -Path .\TestFolder

    Directory: C:\

Path                                    Owner                                   Access
----                                    -----                                   ------
TestFolder                              BUILTIN\Administrators                  BUILTIN\Administrators Allow  FullCo...

Being the PowerShell enthusiast that I am, I modified my command so that only the information I wanted (the path and the owner) was being returned. This is done using the Select-Object cmdlet. Unfortunately, when the command returned those two properties, the Path property was no longer what I was expecting – take a look below. While the example above only returned the name of the folder (TestFolder), I thought I would change this to show the full path (C:\TestFolder) since I was now dealing with a string that included it.

PS C:\> Get-Acl -Path .\TestFolder | Select-Object -Property Path,Owner

Path                                                        Owner
----                                                        -----
Microsoft.PowerShell.Core\FileSystem::C:\TestFolder         BUILTIN\Administrators

The first thing I thought to do was to split the path at the two colons (::) and grab the second element, which I thought would end up being C:\TestFolder. The split method didn’t work so well; here’s what I ended up with.

PS C:\> Get-Acl -Path .\TestFolder | Select-Object -Property @{N='Path';E={($PSItem.Path).Split('::')[-1]}},Owner

Path                                                        Owner
----                                                        -----
\TestFolder                                                 BUILTIN\Administrators

The reason this didn’t work is because the Split() method doesn’t split on each occurrence of two, back-to-back colons like you might expect, it splits on every colon. Since C:\ has a colon, it split there as well. FYI: The use of [-1] returns the last element in an array – good to know, I know. Here’s an example that may help better explain. In this example below, the string is split on every exclamation point (!) and every question mark (?) – not only on the combination of both (!?).

PS C:\> $String = 'Today is the 21st! That is great news, right?'
PS C:\> $String
Today is the 21st! That is great news, right?
PS C:\> $String.Split('!?')
Today is the 21st
 That is great news, right

PS C:\>

What I then decided to do to get this the way I wanted it, was to first replace the two, back-to-back colons with a single character (that was not a part of the string), and then split on that single character. It worked, and here’s what that looks like.

PS C:\> Get-Acl -Path .\TestFolder | Select-Object -Property @{N='Path';E={(($PSItem.Path).Replace('::','@')).Split('@')[-1]}},Owner

Path                                                        Owner
----                                                        -----
C:\TestFolder                                               BUILTIN\Administrators

PS C:\>

It was about this point, that I wanted to see if the -split operator would have handled this the same way and required the additional work that the Replace()/Split() methods did. Of course, after all I did to get this to work how I wanted, I determined I should have started with the -split operator. The -split operator isn’t looking at the characters individually, but instead of, as a whole – two, back-to-back colons is two back-to-back colons.

PS C:\> Get-Acl -Path .\TestFolder | Select-Object -Property @{N='Path';E={(($PSItem.Path) -split '::')[-1]}},Owner

Path                                                        Owner
----                                                        -----
C:\TestFolder                                               BUILTIN\Administrators

A note, if the @{N=…;E={…}} syntax is confusing, or new to you, then spend some time reading this: http://technet.microsoft.com/en-us/library/ff730948.aspx, and then run Get-Help about_Hash_Tables.

This happens just about every time I get to ready to publish a new post. I discovered a better way to handle the problem – much like I did when I considered the -split operator.  It turns out that I could have used a built-in cmdlet to convert the path for me. That’s right, there’s a cmdlet that would have handled everything. Using the Convert-Path cmdlet will convert something like this: Microsoft.PowerShell.Core\FileSystem::C:\TestFolder to this: C:\TestFolder. Here’s the example.

PS C:\> Get-Acl -Path .\TestFolder | Select-Object -Property @{N='Path';E={Convert-Path $PSItem.Path}},Owner

Path                                                        Owner
----                                                        -----
C:\TestFolder                                               BUILTIN\Administrators

PS C:\>

Well, that’s it for this one. As much as I might seem irritated about how I did this three different ways – from the most work to the least – I understand how important this learning process is, and that one day I will be grateful for having gone down this path… (pun intended).