Chapter 1. The Windows PowerShell Interactive Shell

Introduction

Above all else, the design of Windows PowerShell places priority on its use as an efficient and powerful interactive shell. Even its scripting language plays a critical role in this effort, as it too heavily favors interactive use.

What surprises most people when they first launch PowerShell is its similarity to the command prompt that has long existed as part of Windows. Familiar tools continue to run. Familiar commands continue to run. Even familiar hotkeys are the same. Supporting this familiar user interface, though, is a powerful engine that lets you accomplish once cumbersome administrative and scripting tasks with ease.

This chapter introduces PowerShell from the perspective of its interactive shell.

Run Programs, Scripts, and Existing Tools

Problem

You rely on a lot of effort invested in your current tools. You have traditional executables, Perl scripts, VBScript, and of course, a legacy build system that has organically grown into a tangled mess of batch files. You want to use PowerShell, but you don’t want to give up everything you already have.

Solution

To run a program, script, batch file, or other executable command in the system’s path, enter its filename. For these executable types, the extension is optional:

Program.exe arguments
ScriptName.ps1 arguments
BatchFile.cmd arguments

To run a command that contains a space in its name, enclose its filename in single-quotes (') and precede the command with an ampersand (&), known in PowerShell as the invoke operator:

& 'C:\Program Files\Program\Program.exe' arguments

To run a command in the current directory, place .\ in front of its filename:

.\Program.exe arguments

To run a command with spaces in its name from the current directory, precede it with both an ampersand and .\:

& '.\Program With Spaces.exe' arguments

Discussion

In this case, the solution is mainly to use your current tools as you always have. The only difference is that you run them in the PowerShell interactive shell rather than cmd.exe.

Specifying the command name

The final three tips in the Solution merit special attention. They are the features of PowerShell that many new users stumble on when it comes to running programs. The first is running commands that contain spaces. In cmd.exe, the way to run a command that contains spaces is to surround it with quotes:

"C:\Program Files\Program\Program.exe"

In PowerShell, though, placing text inside quotes is part of a feature that lets you evaluate complex expressions at the prompt, as shown in Example 1-1.

Example 1-1. Evaluating expressions at the PowerShell prompt

PS > 1 + 1
2
PS > 26 * 1.15
29.9
PS > "Hello" + " World"
Hello World
PS > "Hello World"
Hello World
PS > "C:\Program Files\Program\Program.exe"
C:\Program Files\Program\Program.exe
PS >

So, a program name in quotes is no different from any other string in quotes. It’s just an expression. As shown previously, the way to run a command in a string is to precede that string with the invoke operator (&). If the command you want to run is a batch file that modifies its environment, see Program: Retain Changes to Environment Variables Set by a Batch File.

Note

By default, PowerShell’s security policies prevent scripts from running. Once you begin writing or using scripts, though, you should configure this policy to something less restrictive. For information on how to configure your execution policy, see Enable Scripting Through an Execution Policy.

The second command that new users (and seasoned veterans before coffee!) sometimes stumble on is running commands from the current directory. In cmd.exe, the current directory is considered part of the path: the list of directories that Windows searches to find the program name you typed. If you are in the C:\Programs directory, cmd.exe looks in C:\Programs (among other places) for applications to run.

PowerShell, like most Unix shells, requires that you explicitly state your desire to run a program from the current directory. To do that, you use the .\Program.exe syntax, as shown previously. This prevents malicious users on your system from littering your hard drive with evil programs that have names similar to (or the same as) commands you might run while visiting that directory.

To save themselves from having to type the location of commonly used scripts and programs, many users put commonly used utilities along with their PowerShell scripts in a “tools” directory, which they add to their system’s path. If PowerShell can find a script or utility in your system’s path, you do not need to explicitly specify its location.

If you want PowerShell to automatically look in your current working directory for scripts, you can add a period (.) to your PATH environment variable.

For more information about updating your system path, see Modify the User or System Path.

If you want to capture the output of a command, you can either save the results into a variable, or save the results into a file. To save the results into a variable, see Store Information in Variables. To save the results into a file, see Store the Output of a Command into a File.

Specifying command arguments

To specify arguments to a command, you can again type them just as you would in other shells. For example, to make a specified file read-only (two arguments to attrib.exe), simply type:

attrib +R c:\path\to\file.txt

Where many scripters get misled when it comes to command arguments is how to change them within your scripts. For example, how do you get the filename from a PowerShell variable? The answer is to define a variable to hold the argument value, and just use that in the place you used to write the command argument:

$filename = "c:\path\to\other\file.txt"
attrib +R $filename

You can use the same technique when you call a PowerShell cmdlet, script, or function:

$filename = "c:\path\to\other\file.txt"
Get-Acl -Path $filename

If you see a solution that uses the Invoke-Expression cmdlet to compose command arguments, it is almost certainly incorrect. The Invoke-Expression cmdlet takes the string that you give it and treats it like a full PowerShell script. As just one example of the problems this can cause, consider the following: filenames are allowed to contain the semicolon (;) character, but when Invoke-Expression sees a semicolon, it assumes that it is a new line of PowerShell script. For example, try running this:

$filename = "c:\file.txt; Write-Warning 'This could be bad'"
Invoke-Expression "Get-Acl -Path $filename"

Given that these dynamic arguments often come from user input, using Invoke-Expression to compose commands can (at best) cause unpredictable script results. Worse, it could result in damage to your system or a security vulnerability.

In addition to letting you supply arguments through variables one at a time, PowerShell also lets you supply several of them at once through a technique known as splatting. For more information about splatting, see Dynamically Compose Command Parameters.

Run a PowerShell Command

Problem

You want to run a PowerShell command.

Solution

To run a PowerShell command, type its name at the command prompt. For example:

PS > Get-Process

Handles  NPM(K)    PM(K)    WS(K)   VM(M)  CPU(s)    Id ProcessName
-------  ------    -----    -----   -----  ------    -- ----------- 
    133       5    11760     7668      46          1112 audiodg
    184       5    33248      508      93          1692 avgamsvr
    143       7    31852      984      97          1788 avgemc

Discussion

The Get-Process command is an example of a native PowerShell command, called a cmdlet. As compared to traditional commands, cmdlets provide significant benefits to both administrators and developers:

  • They share a common and regular command-line syntax.

  • They support rich pipeline scenarios (using the output of one command as the input of another).

  • They produce easily manageable object-based output, rather than error-prone plain-text output.

Because the Get-Process cmdlet generates rich object-based output, you can use its output for many process-related tasks.

Every PowerShell command lets you provide input to the command through its parameters. For more information on providing input to commands, see Running Commands.

The Get-Process cmdlet is just one of the many that PowerShell supports. See Find a Command to Accomplish a Task to learn techniques for finding additional commands that PowerShell supports.

For more information about working with classes from the .NET Framework, see Work with .NET Objects.

Resolve Errors Calling Native Executables

Problem

You have a command line that works from cmd.exe, and want to resolve errors that occur from running that command in PowerShell.

Solution

Enclose any affected command arguments in single quotes to prevent them from being interpreted by PowerShell, and replace any single quotes in the command with two single quotes.

PS > cmd /c echo '!"#$%&''()*+,-./09:;<=>?@AZ[\]^_`az{|}~'
!"#$%&'()*+,-./09:;<=>?@AZ[\]^_`az{|}~

For complicated commands where this does not work, use the verbatim argument (--%) syntax.

PS > cmd /c echo 'quotes' "and" $variables @{ etc = $true }
quotes and System.Collections.Hashtable

PS > cmd --% /c echo 'quotes' "and" $variables @{ etc = $true }
'quotes' "and" $variables @{ etc = $true }

Discussion

One of PowerShell’s primary goals has always been command consistency. Because of this, cmdlets are very regular in the way that they accept parameters. Native executables write their own parameter parsing, so you never know what to expect when working with them. In addition, PowerShell offers many features that make you more efficient at the command line: command substitution, variable expansion, and more. Since many native executables were written before PowerShell was developed, they may use special characters that conflict with these features.

As an example, the command given in the Solution uses all the special characters available on a typical keyboard. Without the quotes, PowerShell treats some of them as language features, as shown in Table 1-1.

Table 1-1. Sample of special characters

Special characterMeaning

"

The beginning (or end) of quoted text

#

The beginning of a comment

$

The beginning of a variable

&

Reserved for future use

( )

Parentheses used for subexpressions

;

Statement separator

{ }

Script block

|

Pipeline separator

`

Escape character

When surrounded by single quotes, PowerShell accepts these characters as written, without the special meaning.

Despite these precautions, you may still sometimes run into a command that doesn’t seem to work when called from PowerShell. For the most part, these can be resolved by reviewing what PowerShell passes to the command and escaping the special characters.

To see exactly what PowerShell passes to that command, you can view the output of the trace source called NativeCommandParameterBinder:

PS > Trace-Command NativeCommandParameterBinder {
    cmd /c echo '!"#$%&''()*+,-./09:;<=>?@AZ[\]^_`az{|}~'
} -PsHost

DEBUG: NativeCommandParameterBinder Information: 0 :  WriteLine
Argument 0: /c
DEBUG: NativeCommandParameterBinder Information: 0 :  WriteLine
Argument 1: echo
DEBUG: NativeCommandParameterBinder Information: 0 :  WriteLine
Argument 2: !#$%&'()*+,-./09:;<=>?@AZ[\]^_`az{|}~
!"#$%&'()*+,-./09:;<=>?@AZ[\]^_`az{|}~

If the command arguments shown in this output don’t match the arguments you expect, they have special meaning to PowerShell and should be escaped.

For a complex enough command that “just used to work,” though, escaping special characters is tiresome. To escape the whole command invocation, use the verbatim argument marker (--%) to prevent PowerShell from interpreting any of the remaining characters on the line. You can place this marker anywhere in the command’s arguments, letting you benefit from PowerShell constructs where appropriate. The following example uses a PowerShell variable for some of the command arguments, but then uses verbatim arguments for the rest:

PS > $username = "Lee"
PS > cmd /c echo Hello $username with 'quotes' "and" $variables @{ etc = $true }
Hello Lee with quotes and System.Collections.Hashtable
PS > cmd /c echo Hello $username `
     --% with 'quotes' "and" $variables @{ etc = $true }
Hello Lee with 'quotes' "and" $variables @{ etc = $true }

While in this mode, PowerShell also accepts cmd.exe-style environment variables—as these are frequently used in commands that “just used to work”:

PS > $env:host = "myhost"
PS > ping %host%
Ping request could not find host %host%. Please check the name and try again.
PS > ping --% %host%

Pinging myhost [127.0.1.1] with 32 bytes of data:
(...)

See Also

Appendix A

Supply Default Values for Parameters

Problem

You want to define a default value for a parameter in a PowerShell command.

Solution

Add an entry to the PSDefaultParameterValues hashtable.

PS > Get-Process

Handles  NPM(K)    PM(K)      WS(K) VM(M)   CPU(s)     Id ProcessName
-------  ------    -----      ----- -----   ------     -- -----------
    150      13     9692       9612    39    21.43    996 audiodg
   1013      84    45572      42716   315     1.67   4596 WWAHost
(...)

PS > $PSDefaultParameterValues["Get-Process:ID"] = $pid
PS > Get-Process

Handles  NPM(K)    PM(K)      WS(K) VM(M)   CPU(s)     Id ProcessName
-------  ------    -----      ----- -----   ------     -- -----------
    584      62   132776     157940   985    13.15   9104 powershell_ise

PS > Get-Process -Id 0

Handles  NPM(K)    PM(K)      WS(K) VM(M)   CPU(s)     Id ProcessName
-------  ------    -----      ----- -----   ------     -- -----------
      0       0        0         20     0               0 Idle

Discussion

In PowerShell, many commands (cmdlets and advanced functions) have parameters that let you configure their behavior. For a full description of how to provide input to commands, see Running Commands. Sometimes, though, supplying values for those parameters at each invocation becomes awkward or repetitive.

Until PowerShell version 3, it was the responsibility of each cmdlet author to recognize awkward or repetitive configuration properties and build support for “preference variables” into the cmdlet itself. For example, the Send-MailMessage cmdlet looks for the $PSEmailServer variable if you do not supply a value for its -SmtpServer parameter.

To make this support more consistent and configurable, PowerShell version 3 introduces the PSDefaultParameterValues preference variable. This preference variable is a hashtable. Like all other PowerShell hashtables, entries come in two parts: the key and the value.

Keys in the PSDefaultParameterValues hashtable must match the pattern cmdlet:parameter—that is, a cmdlet name and parameter name, separated by a colon. Either (or both) may use wildcards, and spaces between the command name, colon, and parameter are ignored.

Values for the cmdlet/parameter pairs can be either a simple parameter value (a string, boolean value, integer, etc.) or a script block. Simple parameter values are what you will use most often.

If you need the default value to dynamically change based on what parameter values are provided so far, you can use a script block as the default. When you do so, PowerShell evaluates the script block and uses its result as the default value. If your script block doesn’t return a result, PowerShell doesn’t apply a default value.

When PowerShell invokes your script block, $args[0] contains information about any parameters bound so far: BoundDefaultParameters, BoundParameters, and BoundPositionalParameters. As one example of this, consider providing default values to the -Credential parameter based on the computer being connected to. Here is a function that simply outputs the credential being used:

function RemoteConnector
{
    param(
        [Parameter()]
        $ComputerName,

        [Parameter(Mandatory = $true)]
        $Credential)

    "Connecting as " + $Credential.UserName
}

Now, you can define a credential map:

PS > $credmap = @{}
PS > $credmap["RemoteComputer1"] = Get-Credential
PS > $credmap["RemoteComputer2"] = Get-Credential

Then, create a parameter default for all Credential parameters that looks at the ComputerName bound parameter:

$PSDefaultParameterValues["*:Credential"] = {
    if($args[0].BoundParameters -contains "ComputerName")
    {
        $cred = $credmap[$PSBoundParameters["ComputerName"]]
        if($cred) { $cred }
    }
}

Here is an example of this in use:

PS > RemoteConnector -ComputerName RemoteComputer1
Connecting as UserForRemoteComputer1
PS > RemoteConnector -ComputerName RemoteComputer2
Connecting as UserForRemoteComputer2
PS > RemoteConnector -ComputerName RemoteComputer3

cmdlet RemoteConnector at command pipeline position 1
Supply values for the following parameters:
Credential: (...)

For more information about working with hashtables in PowerShell, see Hashtables (Associative Arrays).

Invoke a Long-Running or Background Command

Problem

You want to invoke a long-running command on a local or remote computer.

Solution

Invoke the command as a Job to have PowerShell run it in the background:

PS > Start-Job { while($true) { Get-Random; Start-Sleep 5 } } -Name Sleeper

Id              Name            State      HasMoreData     Location
--              ----            -----      -----------     --------
1               Sleeper         Running    True            localhost        

PS > Receive-Job Sleeper
671032665
1862308704
PS > Stop-Job Sleeper

Discussion

PowerShell’s job cmdlets provide a consistent way to create and interact with background tasks. In the Solution, we use the Start-Job cmdlet to launch a background job on the local computer. We give it the name of Sleeper, but otherwise we don’t customize much of its execution environment.

In addition to allowing you to customize the job name, the Start-Job cmdlet also lets you launch the job under alternate user credentials or as a 32-bit process (if run originally from a 64-bit process).

Once you have launched a job, you can use the other Job cmdlets to interact with it:

Get-Job

Gets all jobs associated with the current session. In addition, the -Before, -After, -Newest, and -State parameters let you filter jobs based on their state or completion time.

Wait-Job

Waits for a job until it has output ready to be retrieved.

Receive-Job

Retrieves any output the job has generated since the last call to Receive-Job.

Stop-Job

Stops a job.

Remove-Job

Removes a job from the list of active jobs.

Note

In addition to the Start-Job cmdlet, you can also use the -AsJob parameter in many cmdlets to have them perform their tasks in the background. Two of the most useful examples are the Invoke-Command cmdlet (when operating against remote computers) and the set of WMI-related cmdlets.

If your job generates an error, the Receive-Job cmdlet will display it to you when you receive the results, as shown in Example 1-2. If you want to investigate these errors further, the object returned by Get-Job exposes them through the Error property.

Example 1-2. Retrieving errors from a Job

PS > Start-Job -Name ErrorJob { Write-Error Error! }

Id              Name            State      HasMoreData     Location
--              ----            -----      -----------     --------
1               ErrorJob        Running    True            localhost


PS > Receive-Job ErrorJob
Error!
    + CategoryInfo          : NotSpecified: (:) [Write-Error], WriteError
   Exception
    + FullyQualifiedErrorId : Microsoft.PowerShell.Commands.WriteErrorExc
   eption,Microsoft.PowerShell.Commands.WriteErrorCommand

PS > $job = Get-Job ErrorJob
PS > $job | Format-List *

State         : Completed
HasMoreData   : False
StatusMessage :
Location      : localhost
Command       :  Write-Error Error!
JobStateInfo  : Completed
Finished      : System.Threading.ManualResetEvent
InstanceId    : 801e932c-5580-4c8b-af06-ddd1024840b7
Id            : 1
Name          : ErrorJob
ChildJobs     : {Job2}
Output        : {}
Error         : {}
Progress      : {}
Verbose       : {}
Debug         : {}
Warning       : {}

PS > $job.ChildJobs[0] | Format-List *

State         : Completed
StatusMessage :
HasMoreData   : False
Location      : localhost
Runspace      : System.Management.Automation.RemoteRunspace
Command       :  Write-Error Error!
JobStateInfo  : Completed
Finished      : System.Threading.ManualResetEvent
InstanceId    : 60fa85da-448b-49ff-8116-6eae6c3f5006
Id            : 2
Name          : Job2
ChildJobs     : {}
Output        : {}
Error         : {Microsoft.PowerShell.Commands.WriteErrorException,Microso
                ft.PowerShell.Commands.WriteErrorCommand}
Progress      : {}
Verbose       : {}
Debug         : {}
Warning       : {}


PS > $job.ChildJobs[0].Error
Error!
    + CategoryInfo          : NotSpecified: (:) [Write-Error], WriteError
   Exception
    + FullyQualifiedErrorId : Microsoft.PowerShell.Commands.WriteErrorExc
   eption,Microsoft.PowerShell.Commands.WriteErrorCommand

PS >

As this example shows, jobs are sometimes containers for other jobs, called child jobs. Jobs created through the Start-Job cmdlet will always be child jobs attached to a generic container. To access the errors returned by these jobs, you instead access the errors in its first child job (called child job number zero).

In addition to long-running jobs that execute under control of the current PowerShell session, you might want to register and control jobs that run on a schedule, or independently of the current PowerShell session. PowerShell has a handful of commands to let you work with scheduled jobs like this; for more information, see Manage Scheduled Tasks on a Computer.

Program: Monitor a Command for Changes

As thrilling as our lives are, some days are reduced to running a command over and over and over. Did the files finish copying yet? Is the build finished? Is the site still up?

Usually, the answer to these questions comes from running a command, looking at its output, and then deciding whether it meets your criteria. And usually this means just waiting for the output to change, waiting for some text to appear, or waiting for some text to disappear.

Fortunately, Example 1-3 automates this tedious process for you.

Example 1-3. Watch-Command.ps1

##############################################################################
##
## Watch-Command
##
## From Windows PowerShell Cookbook (O'Reilly)
## by Lee Holmes (http://www.leeholmes.com/guide)
##
##############################################################################

<#

.SYNOPSIS

Watches the result of a command invocation, alerting you when the output
either matches a specified string, lacks a specified string, or has simply
changed.

.EXAMPLE

PS > Watch-Command { Get-Process -Name Notepad | Measure } -UntilChanged
Monitors Notepad processes until you start or stop one.

.EXAMPLE

PS > Watch-Command { Get-Process -Name Notepad | Measure } -Until "Count    : 1"
Monitors Notepad processes until there is exactly one open.

.EXAMPLE

PS > Watch-Command {
     Get-Process -Name Notepad | Measure } -While 'Count    : \d\s*\n'
Monitors Notepad processes while there are between 0 and 9 open
(once number after the colon).

#>


[CmdletBinding(DefaultParameterSetName = "Forever")]
param(
    ## The script block to invoke while monitoring
    [Parameter(Mandatory = $true, Position = 0)]
    [ScriptBlock] $ScriptBlock,

    ## The delay, in seconds, between monitoring attempts
    [Parameter()]
    [Double] $DelaySeconds = 1,

    ## Specifies that the alert sound should not be played
    [Parameter()]
    [Switch] $Quiet,

    ## Monitoring continues only while the output of the
    ## command remains the same.
    [Parameter(ParameterSetName = "UntilChanged", Mandatory = $false)]
    [Switch] $UntilChanged,

    ## The regular expression to search for. Monitoring continues
    ## until this expression is found.
    [Parameter(ParameterSetName = "Until", Mandatory = $false)]
    [String] $Until,

    ## The regular expression to search for. Monitoring continues
    ## until this expression is not found.
    [Parameter(ParameterSetName = "While", Mandatory = $false)]
    [String] $While
)

Set-StrictMode -Version 3

$initialOutput = ""

## Start a continuous loop
while($true)
{
    ## Run the provided script block
    $r = & $ScriptBlock

    ## Clear the screen and display the results
    Clear-Host
    $ScriptBlock.ToString().Trim()
    ""
    $textOutput = $r | Out-String
    $textOutput

    ## Remember the initial output, if we haven't
    ## stored it yet
    if(-not $initialOutput)
    {
        $initialOutput = $textOutput
    }

    ## If we are just looking for any change,
    ## see if the text has changed.
    if($UntilChanged)
    {
        if($initialOutput -ne $textOutput)
        {
            break
        }
    }

    ## If we need to ensure some text is found,
    ## break if we didn't find it.
    if($While)
    {
        if($textOutput -notmatch $While)
        {
            break
        }
    }

    ## If we need to wait for some text to be found,
    ## break if we find it.
    if($Until)
    {
        if($textOutput -match $Until)
        {
            break
        }
    }

    ## Delay
    Start-Sleep -Seconds $DelaySeconds
}

## Notify the user
if(-not $Quiet)
{
    [Console]::Beep(1000, 1000)
}

For more information about running scripts, see Run Programs, Scripts, and Existing Tools.

Notify Yourself of Job Completion

Problem

You want to notify yourself when a long-running job completes.

Solution

Use the Register-TemporaryEvent command given in Create a Temporary Event Subscription to register for the event’s StateChanged event:

PS > $job = Start-Job -Name TenSecondSleep { Start-Sleep 10 }
PS > Register-TemporaryEvent $job StateChanged -Action {
     [Console]::Beep(100,100)
     Write-Host "Job #$($sender.Id) ($($sender.Name)) complete."
}

PS > Job #6 (TenSecondSleep) complete.
PS >

Discussion

When a job completes, it raises a StateChanged event to notify subscribers that its state has changed. We can use PowerShell’s event handling cmdlets to register for notifications about this event, but they are not geared toward this type of one-time event handling. To solve that, we use the Register-TemporaryEvent command given in Create a Temporary Event Subscription.

In our example action block in the Solution, we simply emit a beep and write a message saying that the job is complete.

As another option, you can also update your prompt function to highlight jobs that are complete but still have output you haven’t processed:

$psJobs = @(Get-Job -State Completed | ? { $_.HasMoreData })
if($psJobs.Count -gt 0) {
    ($psJobs | Out-String).Trim() | Write-Host -Fore Yellow }

For more information about events and this type of automatic event handling, see Chapter 32.

Customize Your Shell, Profile, and Prompt

Problem

You want to customize PowerShell’s interactive experience with a personalized prompt, aliases, and more.

Solution

When you want to customize aspects of PowerShell, place those customizations in your personal profile script. PowerShell provides easy access to this profile script by storing its location in the $profile variable.

Note

By default, PowerShell’s security policies prevent scripts (including your profile) from running. Once you begin writing scripts, though, you should configure this policy to something less restrictive. For information on how to configure your execution policy, see Enable Scripting Through an Execution Policy.

To create a new profile (and overwrite one if it already exists):

New-Item -type file -force $profile

To edit your profile (in the Integrated Scripting Environment):

ise $profile

To see your profile file:

Get-ChildItem $profile

Once you create a profile script, you can add a function called prompt that returns a string. PowerShell displays the output of this function as your command-line prompt.

function prompt
{
    "PS [$env:COMPUTERNAME] >"
}

This example prompt displays your computer name, and looks like PS [LEE-DESK] >.

You may also find it helpful to add aliases to your profile. Aliases let you refer to common commands by a name that you choose. Personal profile scripts let you automatically define aliases, functions, variables, or any other customizations that you might set interactively from the PowerShell prompt. Aliases are among the most common customizations, as they let you refer to PowerShell commands (and your own scripts) by a name that is easier to type.

Note

If you want to define an alias for a command but also need to modify the parameters to that command, then define a function instead. For more information, see Dynamically Compose Command Parameters.

For example:

Set-Alias new New-Object
Set-Alias iexplore 'C:\Program Files\Internet Explorer\iexplore.exe'

Your changes will become effective once you save your profile and restart PowerShell. Alternatively, you can reload your profile immediately by running this command:

. $profile

Functions are also very common customizations, with the most popular being the prompt function.

Discussion

The Solution discusses three techniques to make useful customizations to your PowerShell environment: aliases, functions, and a hand-tailored prompt. You can (and will often) apply these techniques at any time during your PowerShell session, but your profile script is the standard place to put customizations that you want to apply to every session.

Note

To remove an alias or function, use the Remove-Item cmdlet:

Remove-Item function:\MyCustomFunction
Remove-Item alias:\new

Although the Prompt function returns a simple string, you can also use the function for more complex tasks. For example, many users update their console window title (by changing the $host.UI.RawUI.WindowTitle variable) or use the Write-Host cmdlet to output the prompt in color. If your prompt function handles the screen output itself, it still needs to return a string (for example, a single space) to prevent PowerShell from using its default. If you don’t want this extra space to appear in your prompt, add an extra space at the end of your Write-Host command and return the backspace ("`b") character, as shown in Example 1-4.

Example 1-4. An example PowerShell prompt

##############################################################################
##
## From Windows PowerShell Cookbook (O'Reilly)
## by Lee Holmes (http://www.leeholmes.com/guide)
##
##############################################################################

Set-StrictMode -Version 3

function Prompt
{
    $id = 1
    $historyItem = Get-History -Count 1
    if($historyItem)
    {
        $id = $historyItem.Id + 1
    }

    Write-Host -ForegroundColor DarkGray "`n[$(Get-Location)]"
    Write-Host -NoNewLine "PS:$id > "
    $host.UI.RawUI.WindowTitle = "$(Get-Location)"

    "`b"
}

In addition to showing the current location, this prompt also shows the ID for that command in your history. This lets you locate and invoke past commands with relative ease:

[C:\]
PS:73 >5 * 5
25

[C:\]
PS:74 >1 + 1
2

[C:\]
PS:75 >Invoke-History 73
5 * 5
25

[C:\]
PS:76 >

Although the profile referenced by $profile is the one you will almost always want to use, PowerShell actually supports four separate profile scripts. For further details on these scripts (along with other shell customization options), see Common Customization Points.

Customize PowerShell’s User Input Behavior

Problem

You want to override the way that PowerShell reads input at the prompt.

Solution

Create a PSConsoleHostReadLine function. In that function, process the user input and return the resulting command. Example 1-5 implements a somewhat ridiculous Notepad-based user input mechanism:

Example 1-5. A Notepad-based user input mechanism

function PSConsoleHostReadLine
{
    $inputFile = Join-Path $env:TEMP PSConsoleHostReadLine
    Set-Content $inputFile "PS > "

    ## Notepad opens. Enter your command in it, save the file,
    ## and then exit.
    notepad $inputFile | Out-Null
    $userInput = Get-Content $inputFile
    $resultingCommand = $userInput.Replace("PS >", "")
    $resultingCommand
}

Discussion

When PowerShell first came on the scene, Unix folks were among the first to notice. They’d enjoyed a powerful shell and a vigorous heritage of automation for years—and “when I’m forced to use Windows, PowerShell rocks” is a phrase we’ve heard many times.

This natural uptake was no mistake. There are many on the team who come from a deep Unix background, and similarities to traditional Unix shells were intentional. When coming from a Unix background, though, we still hear the occasional grumble that tab completion feels weird. Ctrl-R doesn’t invoke history search? Tab cycles through matches, rather than lists them? Abhorrent!

In PowerShell versions 1 or 2, there was nothing you could reasonably do to address this. PowerShell reads its input from the console in what is known as Cooked Mode—where the Windows console subsystem handles all the keypresses, fancy F7 menus, and more. When you press Enter or Tab, PowerShell gets the text of what you have typed so far, but that’s it. There is no way for it to know that you had pressed the (Unix-like) Ctrl-R, Ctrl-A, Ctrl-E, or any other keys.

This issue has been resolved in PowerShell version 3 through the PSConsoleHostReadLine function. When you define this method in the PowerShell console host, PowerShell calls that function instead of the Cooked Mode input functionality. And that’s it—the rest is up to you. If you’d like to implement a custom input method, the freedom (and responsibility) is all yours.

Note

A community implementation of a Bash-like PSConsoleHostReadLine function is available here.

For more information about handling keypresses and other forms of user input, see Chapter 13.

Customize PowerShell’s Command Resolution Behavior

Problem

You want to override or customize the command that PowerShell invokes before it is invoked.

Solution

Assign a script block to one or all of the PreCommandLookupAction, PostCommandLookupAction, or CommandNotFoundAction properties of $executionContext.SessionState.InvokeCommand. Example 1-6 enables easy parent directory navigation when you type multiple dots.

Example 1-6. Enabling easy parent path navigation through CommandNotFoundAction

$executionContext.SessionState.InvokeCommand.CommandNotFoundAction = {
    param($CommandName, $CommandLookupEventArgs)

    ## If the command is only dots
    if($CommandName -match '^\.+$')
    {
        ## Associate a new command that should be invoked instead    
        $CommandLookupEventArgs.CommandScriptBlock = {

            ## Count the number of dots, and run "Set-Location .." one
            ## less time.
            for($counter = 0; $counter -lt $CommandName.Length - 1; $counter++)
            {
                Set-Location ..
            }

        ## We call GetNewClosure() so that the reference to $CommandName can
        ## be used in the new command.
        }.GetNewClosure()
    }
}

PS C:\Users\Lee> cd $pshome
PS C:\Windows\System32\WindowsPowerShell\v1.0> ....
PS C:\Windows>

Discussion

When you invoke a command in PowerShell, the engine goes through three distinct phases:

  1. Retrieve the text of the command.

  2. Find the command for that text.

  3. Invoke the command that was found.

In PowerShell version 3, the $executionContext.SessionState.InvokeCommand property now lets you override any of these stages with script blocks to intercept any or all of the PreCommandLookupAction, PostCommandLookupAction, or CommandNotFoundAction stages.

Each script block receives two parameters: the command name, and an object (CommandLookupEventArgs) to control the command lookup behavior. If your handler assigns a script block to the CommandScriptBlock property of the CommandLookupEventArgs or assigns a CommandInfo to the Command property of the CommandLookupEventArgs, PowerShell will use that script block or command, respectively. If your script block sets the StopSearch property to true, PowerShell will do no further command resolution.

PowerShell invokes the PreCommandLookupAction script block when it knows the name of a command (i.e., Get-Process) but hasn’t yet looked for the command itself. You can override this action if you want to react primarily based on the text of the command name or want to preempt PowerShell’s regular command or alias resolution. For example, Example 1-7 demonstrates a PreCommandLookupAction that looks for commands with an asterisk before their name. When it sees one, it enables the -Verbose parameter.

Example 1-7. Customizing the PreCommandLookupAction

$executionContext.SessionState.InvokeCommand.PreCommandLookupAction = {
    param($CommandName, $CommandLookupEventArgs)

    ## If the command name starts with an asterisk, then
    ## enable its Verbose parameter
    if($CommandName -match "\*")
    {
        ## Remove the leading asterisk
        $NewCommandName = $CommandName -replace '\*',''

        ## Create a new script block that invokes the actual command,
        ## passes along all original arguments, and adds in the -Verbose
        ## parameter
        $CommandLookupEventArgs.CommandScriptBlock = {
            & $NewCommandName @args -Verbose

        ## We call GetNewClosure() so that the reference to $NewCommandName
        ## can be used in the new command.
        }.GetNewClosure()
    }
}

PS > dir > 1.txt
PS > dir > 2.txt
PS > del 1.txt
PS > *del 2.txt
VERBOSE: Performing operation "Remove file" on Target "C:\temp\tempfolder\2.txt".

After PowerShell executes the PreCommandLookupAction (if one exists and doesn’t return a command), it goes through its regular command resolution. If it finds a command, it invokes the script block associated with the PostCommandLookupAction. You can override this action if you want to react primarily to a command that is just about to be invoked. Example 1-8 demonstrates a PostCommandLookupAction that tallies the commands you use most frequently.

Example 1-8. Customizing the PostCommandLookupAction

$executionContext.SessionState.InvokeCommand.PostCommandLookupAction = {
    param($CommandName, $CommandLookupEventArgs)

    ## Stores a hashtable of the commands we use most frequently
    if(-not (Test-Path variable:\CommandCount))
    {
        $global:CommandCount = @{}
    }

    ## If it was launched by us (rather than as an internal helper
    ## command), record its invocation.
    if($CommandLookupEventArgs.CommandOrigin -eq "Runspace")
    {
        $commandCount[$CommandName] = 1 + $commandCount[$CommandName]
    }
}

PS > Get-Variable commandCount
PS > Get-Process -id $pid
PS > Get-Process -id $pid
PS > $commandCount

Name                           Value
----                           -----
Out-Default                    4
Get-Variable                   1
prompt                         4
Get-Process                    2

If command resolution is unsuccessful, PowerShell invokes the CommandNotFoundAction script block if one exists. At its simplest, you can override this action if you want to recover from or override PowerShell’s error behavior when it cannot find a command.

As a more advanced application, the CommandNotFoundAction lets you write PowerShell extensions that alter their behavior based on the form of the name, rather than the arguments passed to it. For example, you might want to automatically launch URLs just by typing them or navigate around providers just by typing relative path locations.

The Solution gives an example of implementing this type of handler. While dynamic relative path navigation is not a built-in feature of PowerShell, it is possible to get a very reasonable alternative by intercepting the CommandNotFoundAction. If we see a missing command that has a pattern we want to handle (a series of dots), we return a script block that does the appropriate relative path navigation.

Find a Command to Accomplish a Task

Problem

You want to accomplish a task in PowerShell but don’t know the command or cmdlet to accomplish that task.

Solution

Use the Get-Command cmdlet to search for and investigate commands.

To get the summary information about a specific command, specify the command name as an argument:

Get-Command CommandName

To get the detailed information about a specific command, pipe the output of Get-Command to the Format-List cmdlet:

Get-Command CommandName | Format-List

To search for all commands with a name that contains text, surround the text with asterisk characters:

Get-Command *text*

To search for all commands that use the Get verb, supply Get to the -Verb parameter:

Get-Command -Verb Get

To search for all commands that act on a service, use Service as the value of the -Noun parameter:

Get-Command -Noun Service

Discussion

One of the benefits that PowerShell provides administrators is the consistency of its command names. All PowerShell commands (called cmdlets) follow a regular Verb-Noun pattern—for example, Get-Process, Get-EventLog, and Set-Location. The verbs come from a relatively small set of standard verbs (as listed in Appendix J) and describe what action the cmdlet takes. The nouns are specific to the cmdlet and describe what the cmdlet acts on.

Knowing this philosophy, you can easily learn to work with groups of cmdlets. If you want to start a service on the local machine, the standard verb for that is Start. A good guess would be to first try Start-Service (which in this case would be correct), but typing Get-Command -Verb Start would also be an effective way to see what things you can start. Going the other way, you can see what actions are supported on services by typing Get-Command -Noun Service.

When you use the Get-Command cmdlet, PowerShell returns results from the list of all commands available on your system. If you’d instead like to search just commands from modules that you’ve loaded either explicitly or through autoloading, use the -ListImported parameter. For more information about PowerShell’s autoloading of commands, see Extend Your Shell with Additional Commands.

See Get Help on a Command for a way to list all commands along with a brief description of what they do.

The Get-Command cmdlet is one of the three commands you will use most commonly as you explore Windows PowerShell. The other two commands are Get-Help and Get-Member.

There is one important point to keep in mind when it comes to looking for a PowerShell command to accomplish a particular task. Many times, that PowerShell command does not exist, because the task is best accomplished the same way it always was—for example, ipconfig.exe to get IP configuration information, netstat.exe to list protocol statistics and current TCP/IP network connections, and many more.

For more information about the Get-Command cmdlet, type Get-Help Get-Command.

Get Help on a Command

Problem

You want to learn how a specific command works and how to use it.

Solution

The command that provides help and usage information about a command is called Get-Help. It supports several different views of the help information, depending on your needs.

To get the summary of help information for a specific command, provide the command’s name as an argument to the Get-Help cmdlet. This primarily includes its synopsis, syntax, and detailed description:

Get-Help CommandName

or:

CommandName -?

To get the detailed help information for a specific command, supply the -Detailed flag to the Get-Help cmdlet. In addition to the summary view, this also includes its parameter descriptions and examples:

Get-Help CommandName -Detailed

To get the full help information for a specific command, supply the -Full flag to the Get-Help cmdlet. In addition to the detailed view, this also includes its full parameter descriptions and additional notes:

Get-Help CommandName -Full

To get only the examples for a specific command, supply the -Examples flag to the Get-Help cmdlet:

Get-Help CommandName -Examples

To retrieve the most up-to-date online version of a command’s help topic, supply the -Online flag to the Get-Help cmdlet:

Get-Help CommandName -Online

To view a searchable, graphical view of a help topic, use the -ShowWindow parameter:

Get-Help CommandName -ShowWindow

To find all help topics that contain a given keyword, provide that keyword as an argument to the Get-Help cmdlet. If the keyword isn’t also the name of a specific help topic, this returns all help topics that contain the keyword, including its name, category, and synopsis:

Get-Help Keyword

Discussion

The Get-Help cmdlet is the primary way to interact with the help system in PowerShell. Like the Get-Command cmdlet, the Get-Help cmdlet supports wildcards. If you want to list all commands that have help content that matches a certain pattern (for example, *process*), you can simply type Get-Help *process*.

If the pattern matches only a single command, PowerShell displays the help for that command. Although command wildcarding and keyword searching is a helpful way to search PowerShell help, see Program: Search Help for Text for a script that lets you search the help content for a specified pattern.

While there are thousands of pages of custom-written help content at your disposal, PowerShell by default includes only information that it can automatically generate from the information contained in the commands themselves: names, parameters, syntax, and parameter defaults. You need to update your help content to retrieve the rest. The first time you run Get-Help as an administrator on a system, PowerShell offers to download this updated help content:

PS > Get-Help Get-Process

Do you want to run Update-Help?
The Update-Help cmdlet downloads the newest Help files for Windows
PowerShell modules and installs them on your computer. For more details,
see the help topic at http://go.microsoft.com/fwlink/?LinkId=210614.

[Y] Yes  [N] No  [S] Suspend  [?] Help (default is "Y"):

Answer Y to this prompt, and PowerShell automatically downloads and installs the most recent help content for all modules on your system. For more information on updatable help, see Update System Help Content.

If you’d like to generate a list of all cmdlets and aliases (along with their brief synopses), run the following command:

Get-Help * -Category Cmdlet | Select-Object Name,Synopsis | Format-Table -Auto

In addition to console-based help, PowerShell also offers online access to its help content. The Solution demonstrates how to quickly access online help content.

The Get-Help cmdlet is one of the three commands you will use most commonly as you explore Windows PowerShell. The other two commands are Get-Command and Get-Member.

For more information about the Get-Help cmdlet, type Get-Help Get-Help.

Update System Help Content

Problem

You want to update your system’s help content to the latest available.

Solution

Run the Update-Help command. To retrieve help from a local path, use the -SourcePath cmdlet parameter:

Update-Help

or:

Update-Help -SourcePath \\helpserver\help

Discussion

One of PowerShell’s greatest strengths is the incredible detail of its help content. Counting only the help content and about_* topics that describe core functionality, PowerShell’s help includes approximately half a million words and would span 1,200 pages if printed.

The challenge that every version of PowerShell has been forced to deal with is that this help content is written at the same time as PowerShell itself. Given that its goal is to help the user, the content that’s ready by the time a version of PowerShell releases is a best-effort estimate of what users will need help with.

As users get their hands on PowerShell, they start to have questions. Some of these are addressed by the help topics, while some of them aren’t. Sometimes the help is simply incorrect due to a product change during the release. Before PowerShell version 3, resolving these issues meant waiting for the next release of Windows or relying solely on Get-Help’s -Online parameter. To address this, PowerShell version 3 introduces updatable help.

It’s not only possible to update help, but in fact the Update-Help command is the only way to get help on your system. Out of the box, PowerShell provides an experience derived solely from what is built into the commands themselves: name, syntax, parameters, and default values.

The first time you run Get-Help as an administrator on a system, PowerShell offers to download updated help content:

PS > Get-Help Get-Process

Do you want to run Update-Help?
The Update-Help cmdlet downloads the newest Help files for Windows
PowerShell modules and installs them on your computer. For more details,
see the help topic at http://go.microsoft.com/fwlink/?LinkId=210614.

[Y] Yes  [N] No  [S] Suspend  [?] Help (default is "Y"):

Answer Y to this prompt, and PowerShell automatically downloads and installs the most recent help content for all modules on your system.

Note

If you are building a system image and want to prevent this prompt from ever appearing, set the registry key HKLM:\Software\Microsoft\PowerShell\DisablePromptToUpdateHelp to 1.

In addition to the prompt-driven experience, you can call the Update-Help cmdlet directly.

Both experiences look at each module on your system, comparing the help you have for that module with the latest version online. For in-box modules, PowerShell uses download.microsoft.com to retrieve updated help content. Other modules that you download from the Internet can use the HelpInfoUri module key to support their own updatable help.

By default, the Update-Help command retrieves its content from the Internet. If you want to update help on a machine not connected to the Internet, you can use the -SourcePath parameter of the Update-Help cmdlet. This path represents a directory or UNC path where PowerShell should look for updated help content. To populate this content, first use the Save-Help cmdlet to download the files, and then copy them to the source location.

For more information about PowerShell help, see Get Help on a Command.

Program: Search Help for Text

Both the Get-Command and Get-Help cmdlets let you search for command names that match a given pattern. However, when you don’t know exactly what portions of a command name you are looking for, you will more often have success searching through the help content for an answer. On Unix systems, this command is called Apropos.

The Get-Help cmdlet automatically searches the help database for keyword references when it can’t find a help topic for the argument you supply. In addition to that, you might want to extend this even further to search for text patterns or even help topics that talk about existing help topics. PowerShell’s help facilities support a version of wildcarded content searches, but don’t support full regular expressions.

That doesn’t need to stop us, though, as we can write the functionality ourselves.

To run this program, supply a search string to the Search-Help script (given in Example 1-9). The search string can be either simple text or a regular expression. The script then displays the name and synopsis of all help topics that match. To see the help content for that topic, use the Get-Help cmdlet.

Example 1-9. Search-Help.ps1

##############################################################################
##
## Search-Help
##
## From Windows PowerShell Cookbook (O'Reilly)
## by Lee Holmes (http://www.leeholmes.com/guide)
##
##############################################################################

<#

.SYNOPSIS

Search the PowerShell help documentation for a given keyword or regular
expression. For simple keyword searches in PowerShell version two or three,
simply use "Get-Help <keyword>"

.EXAMPLE

PS > Search-Help hashtable
Searches help for the term 'hashtable'

.EXAMPLE

PS > Search-Help "(datetime|ticks)"
Searches help for the term datetime or ticks, using the regular expression
syntax.

#>

param(
    ## The pattern to search for
    [Parameter(Mandatory = $true)]
    $Pattern
)

$helpNames = $(Get-Help * | Where-Object { $_.Category -ne "Alias" })

## Go through all of the help topics
foreach($helpTopic in $helpNames)
{
    ## Get their text content, and
    $content = Get-Help -Full $helpTopic.Name | Out-String
    if($content -match "(.{0,30}$pattern.{0,30})")
    {
        $helpTopic | Add-Member NoteProperty Match $matches[0].Trim()
        $helpTopic | Select-Object Name,Match
    }
}

For more information about running scripts, see Run Programs, Scripts, and Existing Tools.

Launch PowerShell at a Specific Location

Problem

You want to launch a PowerShell session in a specific location.

Solution

Both Windows and PowerShell offer several ways to launch PowerShell in a specific location:

  • Explorer’s address bar

  • PowerShell’s command-line arguments

  • Community extensions

Discussion

If you are browsing the filesystem with Windows Explorer, typing PowerShell into the address bar launches PowerShell in that location (as shown in Figure 1-1).

Launching PowerShell from Windows Explorer

Figure 1-1. Launching PowerShell from Windows Explorer

Additionally, Windows 8 offers an Open Windows PowerShell option directly from the File menu, as shown in Figure 1-2).

For another way to launch PowerShell from Windows Explorer, several members of the PowerShell community have written power toys and Windows Explorer extensions that provide a “Launch PowerShell Here” option when you right-click on a folder from Windows Explorer. An Internet search for “PowerShell Here” turns up several.

If you aren’t browsing the desired folder with Windows Explorer, you can use StartRun (or any other means of launching an application) to launch PowerShell at a specific location. For that, use PowerShell’s -NoExit parameter, along with the implied -Command parameter. In the -Command parameter, call the Set-Location cmdlet to initially move to your desired location.

PowerShell -NoExit Set-Location 'C:\Program Files'
Launching PowerShell in Windows 8

Figure 1-2. Launching PowerShell in Windows 8

Invoke a PowerShell Command or Script from Outside PowerShell

Problem

You want to invoke a PowerShell command or script from a batch file, a logon script, a scheduled task, or any other non-PowerShell application.

Solution

To invoke a PowerShell command, use the -Command parameter:

PowerShell -Command Get-Process; Read-Host

To launch a PowerShell script, use the -File parameter:

PowerShell -File 'full path to script' arguments

For example:

PowerShell -File 'c:\shared scripts\Get-Report.ps1' Hello World

Discussion

By default, any arguments to PowerShell.exe get interpreted as commands to run. PowerShell runs the command as though you had typed it in the interactive shell, and then exits. You can customize this behavior by supplying other parameters to PowerShell.exe, such as -NoExit, -NoProfile, and more.

Note

If you are the author of a program that needs to run PowerShell scripts or commands, PowerShell lets you call these scripts and commands much more easily than calling its command-line interface. For more information about this approach, see Add PowerShell Scripting to Your Own Program.

Since launching a script is so common, PowerShell provides the -File parameter to eliminate the complexities that arise from having to invoke a script from the -Command parameter. This technique lets you invoke a PowerShell script as the target of a logon script, advanced file association, scheduled task, and more.

Note

When PowerShell detects that its input or output streams have been redirected, it suppresses any prompts that it might normally display. If you want to host an interactive PowerShell prompt inside another application (such as Emacs), use - as the argument for the -File parameter. In PowerShell (as with traditional Unix shells), this implies “taken from standard input.”

powershell -File -

If the script is for background automation or a scheduled task, these scripts can sometimes interfere with (or become influenced by) the user’s environment. For these situations, three parameters come in handy:

-NoProfile

Runs the command or script without loading user profile scripts. This makes the script launch faster, but it primarily prevents user preferences (e.g., aliases and preference variables) from interfering with the script’s working environment.

-WindowStyle

Runs the command or script with the specified window style—most commonly Hidden. When run with a window style of Hidden, PowerShell hides its main window immediately. For more ways to control the window style from within PowerShell, see Launch a Process.

-ExecutionPolicy

Runs the command or script with a specified execution policy applied only to this instance of PowerShell. This lets you write PowerShell scripts to manage a system without having to change the system-wide execution policy. For more information about scoped execution policies, see Enable Scripting Through an Execution Policy.

If the arguments to the -Command parameter become complex, special character handling in the application calling PowerShell (such as cmd.exe) might interfere with the command you want to send to PowerShell. For this situation, PowerShell supports an EncodedCommand parameter: a Base64-encoded representation of the Unicode string you want to run. Example 1-10 demonstrates how to convert a string containing PowerShell commands to a Base64-encoded form.

Example 1-10. Converting PowerShell commands into a Base64-encoded form

$commands = '1..10 | % { "PowerShell Rocks" }'
$bytes = [System.Text.Encoding]::Unicode.GetBytes($commands)
$encodedString = [Convert]::ToBase64String($bytes)

Once you have the encoded string, you can use it as the value of the EncodedCommand parameter, as shown in Example 1-11.

Example 1-11. Launching PowerShell with an encoded command from cmd.exe

Microsoft Windows [Version 6.0.6000]
Copyright (c) 2006 Microsoft Corporation. All rights reserved.

C:\Users\Lee>PowerShell -EncodedCommand MQAuAC4AMQAwACAAfAAgACUAIAB7ACAAIgBQAG8A
   dwBlAHIAUwBoAGUAbABsACAAUgBvAGMAawBzACIAIAB9AA==
PowerShell Rocks
PowerShell Rocks
PowerShell Rocks
PowerShell Rocks
PowerShell Rocks
PowerShell Rocks
PowerShell Rocks
PowerShell Rocks
PowerShell Rocks
PowerShell Rocks

For more information about running scripts, see Run Programs, Scripts, and Existing Tools.

Understand and Customize PowerShell’s Tab Completion

Problem

You want to customize how PowerShell reacts to presses of the Tab key (and additionally, Ctrl-Space in the case of IntelliSense in the Integrated Scripting Environment).

Solution

Create a custom function called TabExpansion2. PowerShell invokes this function when you press Tab, or when it invokes IntelliSense in the Integrated Scripting Environment.

Discussion

When you press Tab, PowerShell invokes a facility known as tab expansion: replacing what you’ve typed so far with an expanded version of that (if any apply.) For example, if you type Set-Location C:\ and then press Tab, PowerShell starts cycling through directories under C:\ for you to navigate into.

The features offered by PowerShell’s built-in tab expansion are quite rich, as shown in Table 1-2.

Table 1-2. Tab expansion features in Windows PowerShell

Description

Example

Command completion. Completes command names when current text appears to represent a command invocation.

Get-Ch <Tab>

Parameter completion. Completes command parameters for the current command.

Get-ChildItem -Pat <Tab>

Argument completion. Completes command arguments for the current command parameter. This applies to any command argument that takes a fixed set of values (enumerations or parameters that define a ValidateSet attribute). In addition, PowerShell contains extended argument completion for module names, help topics, CIM / WMI classes, event log names, job IDs and names, process IDs and names, provider names, drive names, service names and display names, and trace source names.

Set-ExecutionPolicy -ExecutionPolicy <Tab>

History text completion. Replaces the current input with items from the command history that match the text after the # character.

# Process <Tab>

History ID completion. Replaces the current input with the command line from item number ID in your command history.

# 12 <Tab>

Filename completion. Replaces the current parameter value with file names that match what you’ve typed so far. When applied to the Set-Location cmdlet, PowerShell further filters results to only directories.

Set-Location C:\Windows\S<Tab>

Operator completion. Replaces the current text with a matching operator. This includes flags supplied to the switch statement.

"Hello World" -rep <Tab>

switch - c <Tab>

Variable completion. Replaces the current text with available PowerShell variables. In the Integrated Scripting Environment, PowerShell incorporates variables even from script content that has never been invoked.

$myGreeting = "Hello World"; $myGr <Tab>

Member completion. Replaces member names for the currently referenced variable or type. When PowerShell can infer the members from previous commands in the pipeline, it even supports member completion within script blocks.

[Console]::Ba<TAB>

Get-Process | Where-Object { $_.Ha <Tab>

Type completion. Replaces abbreviated type names with their namespace-qualified name.

[PSSer<TAB>

$l = New-Object List[Stri <Tab>

If you want to extend PowerShell’s tab expansion capabilities, define a function called TabExpansion2. You can add this to your PowerShell profile directly, or dot-source it from your profile. Example 1-12 demonstrates an example custom tab expansion function that extends the functionality already built into PowerShell.

Example 1-12. A sample implementation of TabExpansion2

##############################################################################
##
## TabExpansion2
##
## From Windows PowerShell Cookbook (O'Reilly)
## by Lee Holmes (http://www.leeholmes.com/guide)
##
##############################################################################

function TabExpansion2
{
    [CmdletBinding(DefaultParameterSetName = 'ScriptInputSet')]
    Param(
        [Parameter(ParameterSetName = 'ScriptInputSet', Mandatory = $true,
            Position = 0)]
        [string] $inputScript,
    
        [Parameter(ParameterSetName = 'ScriptInputSet', Mandatory = $true,
            Position = 1)]
        [int] $cursorColumn,

        [Parameter(ParameterSetName = 'AstInputSet', Mandatory = $true,
            Position = 0)]
        [System.Management.Automation.Language.Ast] $ast,

        [Parameter(ParameterSetName = 'AstInputSet', Mandatory = $true,
            Position = 1)]
        [System.Management.Automation.Language.Token[]] $tokens,

        [Parameter(ParameterSetName = 'AstInputSet', Mandatory = $true,
            Position = 2)]
        [System.Management.Automation.Language.IScriptPosition] $positionOfCursor,
    
        [Parameter(ParameterSetName = 'ScriptInputSet', Position = 2)]
        [Parameter(ParameterSetName = 'AstInputSet', Position = 3)]
        [Hashtable] $options = $null
    )

    End
    {
        ## Create a new 'Options' hashtable if one has not been supplied.
        ## In this hashtable, you can add keys for the following options, using
        ## $true or $false for their values:
        ##
        ## IgnoreHiddenShares - Ignore hidden UNC shares (such as \\COMPUTER\ADMIN$)
        ## RelativePaths - When expanding filenames and paths, $true forces PowerShell
        ## to replace paths with relative paths. When $false, forces PowerShell to
        ## replace them with absolute paths. By default, PowerShell makes this
        ## decision based on what you had typed so far before invoking tab completion.
        ## LiteralPaths - Prevents PowerShell from replacing special file characters
        ## (such as square brackets and back-ticks) with their escaped equivalent.
        if(-not $options) { $options = @{} }

        ## Demonstrate some custom tab expansion completers for parameters.
        ## This is a hashtable of parameter names (and optionally cmdlet names)
        ## that we add to the $options hashtable.
        ##
        ## When PowerShell evaluates the script block, $args gets the
        ## following: command name, parameter, word being completed,
        ## AST of the command being completed, and currently bound arguments.
        $options["CustomArgumentCompleters"] = @{
            "Get-ChildItem:Filter" = { "*.ps1","*.txt","*.doc" }
            "ComputerName" = { "ComputerName1","ComputerName2","ComputerName3" }
        }

        ## Also define a completer for a native executable.
        ## When PowerShell evaluates the script block, $args gets the
        ## word being completed, and AST of the command being completed.
        $options["NativeArgumentCompleters"] = @{
            "attrib" = { "+R","+H","+S" }
        }

        ## Define a "quick completions" list that we'll cycle through
        ## when the user types '!!' followed by TAB.
        $quickCompletions = @(
            'Get-Process -Name PowerShell | ? Id -ne $pid | Stop-Process',
            'Set-Location $pshome',
            ('$errors = $error | % { $_.InvocationInfo.Line }; Get-History | ' +
                ' ? { $_.CommandLine -notin $errors }')
        )

        ## First, check the built-in tab completion results
        $result = $null
        if ($psCmdlet.ParameterSetName -eq 'ScriptInputSet')
        {
            $result = [Management.Automation.CommandCompletion]::CompleteInput(
                <#inputScript#>  $inputScript,
                <#cursorColumn#> $cursorColumn,
                <#options#>      $options)
        }
        else
        {
            $result = [Management.Automation.CommandCompletion]::CompleteInput(
                <#ast#>              $ast,
                <#tokens#>           $tokens,
                <#positionOfCursor#> $positionOfCursor,
                <#options#>          $options)
        }

        ## If we didn't get a result
        if($result.CompletionMatches.Count -eq 0)
        {
            ## If this was done at the command-line or in a remote session,
            ## create an AST out of the input
            if ($psCmdlet.ParameterSetName -eq 'ScriptInputSet')
            {
                $ast = [System.Management.Automation.Language.Parser]::ParseInput(
                    $inputScript, [ref]$tokens, [ref]$null)
            }

            ## In this simple example, look at the text being supplied.
            ## We could do advanced analysis of the AST here if we wanted,
            ## but in this case just use its text. We use a regular expression
            ## to check if the text started with two exclamations, and then
            ## use a match group to retain the rest.
            $text = $ast.Extent.Text
            if($text -match '^!!(.*)')
            {
                ## Extract the rest of the text from the regular expression
                ## match group.
                $currentCompletionText = $matches[1].Trim()

                ## Go through each of our quick completions and add them to
                ## our completion results. The arguments to the completion results
                ## are the text to be used in tab completion, a potentially shorter
                ## version to use for display (i.e., IntelliSense in the ISE),
                ## the type of match, and a potentially more verbose description to
                ## be used as a tool tip.
                $quickCompletions |
                    Where-Object { $_ -match $currentCompletionText } |
                    Foreach-Object { $result.CompletionMatches.Add(
                (New-Object Management.Automation.CompletionResult $_,$_,"Text",$_) )
                }
            }
        }

        return $result
    }     
}

Program: Learn Aliases for Common Commands

In interactive use, full cmdlet names (such as Get-ChildItem) are cumbersome and slow to type. Although aliases are much more efficient, it takes a while to discover them. To learn aliases more easily, you can modify your prompt to remind you of the shorter version of any aliased commands that you use.

This involves two steps:

  1. Add the program Get-AliasSuggestion.ps1, shown in Example 1-13, to your tools directory or another directory.

    Example 1-13. Get-AliasSuggestion.ps1

    ##############################################################################
    ##
    ## Get-AliasSuggestion
    ##
    ## From Windows PowerShell Cookbook (O'Reilly)
    ## by Lee Holmes (http://www.leeholmes.com/guide)
    ##
    ##############################################################################
    
    <#
    
    .SYNOPSIS
    
    Get an alias suggestion from the full text of the last command. Intended to
    be added to your prompt function to help learn aliases for commands.
    
    .EXAMPLE
    
    PS > Get-AliasSuggestion Remove-ItemProperty
    Suggestion: An alias for Remove-ItemProperty is rp
    
    #>
    
    param(
        ## The full text of the last command
        $LastCommand
    )
    
    Set-StrictMode -Version 3
    
    $helpMatches = @()
    
    ## Find all of the commands in their last input
    $tokens = [Management.Automation.PSParser]::Tokenize(
        $lastCommand, [ref] $null)
    $commands = $tokens | Where-Object { $_.Type -eq "Command" }
    
    ## Go through each command
    foreach($command in $commands)
    {
        ## Get the alias suggestions
        foreach($alias in Get-Alias -Definition $command.Content)
        {
            $helpMatches += "Suggestion: An alias for " +
                "$($alias.Definition) is $($alias.Name)"
        }
    }
    
    $helpMatches
  2. Add the text from Example 1-14 to the Prompt function in your profile. If you do not yet have a Prompt function, see Customize Your Shell, Profile, and Prompt to learn how to add one.

    Example 1-14. A useful prompt to teach you aliases for common commands

    function Prompt
    {
        ## Get the last item from the history
        $historyItem = Get-History -Count 1
    
        ## If there were any history items
        if($historyItem)
        {
            ## Get the training suggestion for that item
            $suggestions = @(Get-AliasSuggestion $historyItem.CommandLine)
            ## If there were any suggestions
            if($suggestions)
            {
                ## For each suggestion, write it to the screen
                foreach($aliasSuggestion in $suggestions)
                {
                    Write-Host "$aliasSuggestion"
                }
                Write-Host ""
    
            }
        }
    
        ## Rest of prompt goes here
        "PS [$env:COMPUTERNAME] >"
    }

For more information about running scripts, see Run Programs, Scripts, and Existing Tools.

Program: Learn Aliases for Common Parameters

Problem

You want to learn aliases defined for command parameters.

Solution

Use the Get-ParameterAlias script, as shown in Example 1-15, to return all aliases for parameters used by the previous command in your session history.

Example 1-15. Get-ParameterAlias.ps1

##############################################################################
##
## Get-ParameterAlias
##
## From Windows PowerShell Cookbook (O'Reilly)
## by Lee Holmes (http://www.leeholmes.com/guide)
##
##############################################################################

<#

.SYNOPSIS

Looks in the session history, and returns any aliases that apply to
parameters of commands that were used.

.EXAMPLE

PS > dir -ErrorAction SilentlyContinue
PS > Get-ParameterAlias
An alias for the 'ErrorAction' parameter of 'dir' is ea

#>

Set-StrictMode -Version 3

## Get the last item from their session history
$history = Get-History -Count 1
if(-not $history)
{
    return
}

## And extract the actual command line they typed
$lastCommand = $history.CommandLine

## Use the Tokenizer API to determine which portions represent
## commands and parameters to those commands
$tokens = [System.Management.Automation.PsParser]::Tokenize(
    $lastCommand, [ref] $null)
$currentCommand = $null

## Now go through each resulting token
foreach($token in $tokens)
{
    ## If we've found a new command, store that.
    if($token.Type -eq "Command")
    {
        $currentCommand = $token.Content
    }

    ## If we've found a command parameter, start looking for aliases
    if(($token.Type -eq "CommandParameter") -and ($currentCommand))
    {
        ## Remove the leading "-" from the parameter
        $currentParameter = $token.Content.TrimStart("-")

        ## Determine all of the parameters for the current command.
        (Get-Command $currentCommand).Parameters.GetEnumerator() |

            ## For parameters that start with the current parameter name,
            Where-Object { $_.Key -like "$currentParameter*" } |

            ## return all of the aliases that apply. We use "starts with"
            ## because the user might have typed a shortened form of
            ## the parameter name.
            Foreach-Object {
                $_.Value.Aliases | Foreach-Object {
                    "Suggestion: An alias for the '$currentParameter' " +
                    "parameter of '$currentCommand' is '$_'"
                }
            }
    }
}

Discussion

To make it easy to type command parameters, PowerShell lets you type only as much of the command parameter as is required to disambiguate it from other parameters of that command. In addition to shortening implicitly supported by the shell, cmdlet authors can also define explicit aliases for their parameters—for example, CN as a short form for ComputerName.

While helpful, these aliases are difficult to discover.

If you want to see the aliases for a specific command, you can access its Parameters collection:

PS > (Get-Command New-TimeSpan).Parameters.Values | Select Name,Aliases

Name                                         Aliases
----                                         -------
Start                                        {LastWriteTime}
End                                          {}
Days                                         {}
Hours                                        {}
Minutes                                      {}
Seconds                                      {}
Verbose                                      {vb}
Debug                                        {db}
ErrorAction                                  {ea}
WarningAction                                {wa}
ErrorVariable                                {ev}
WarningVariable                              {wv}
OutVariable                                  {ov}
OutBuffer                                    {ob}

If you want to learn any aliases for parameters in your previous command, simply run Get-ParameterAlias.ps1. To make PowerShell do this automatically, add a call to Get-ParameterAlias.ps1 in your prompt.

This script builds on two main features: PowerShell’s Tokenizer API, and the rich information returned by the Get-Command cmdlet. PowerShell’s Tokenizer API examines its input and returns PowerShell’s interpretation of the input: commands, parameters, parameter values, operators, and more. Like the rich output produced by most of PowerShell’s commands, Get-Command returns information about a command’s parameters, parameter sets, output type (if specified), and more.

For more information about the Tokenizer API, see Parse and Interpret PowerShell Scripts.

Access and Manage Your Console History

Problem

After working in the shell for a while, you want to invoke commands from your history, view your command history, and save your command history.

Solution

The shortcuts given in Customize Your Shell, Profile, and Prompt let you manage your history, but PowerShell offers several features to help you work with your console in even more detail.

To get the most recent commands from your session, use the Get-History cmdlet (or its alias of h):

Get-History

To rerun a specific command from your session history, provide its ID to the Invoke-History cmdlet (or its alias of ihy):

Invoke-History ID

To increase (or limit) the number of commands stored in your session history, assign a new value to the $MaximumHistoryCount variable:

$MaximumHistoryCount = Count

To save your command history to a file, pipe the output of Get-History to the Export-CliXml cmdlet:

Get-History | Export-CliXml Filename

To add a previously saved command history to your current session history, call the Import-CliXml cmdlet and then pipe that output to the Add-History cmdlet:

Import-CliXml Filename | Add-History

To clear all commands from your session history, use the Clear-History cmdlet:

Clear-History

Discussion

Unlike the console history hotkeys discussed in Customize Your Shell, Profile, and Prompt, the Get-History cmdlet produces rich objects that represent information about items in your history. Each object contains that item’s ID, command line, start of execution time, and end of execution time.

Once you know the ID of a history item (as shown in the output of Get-History), you can pass it to Invoke-History to execute that command again. The example prompt function shown in Customize Your Shell, Profile, and Prompt makes working with prior history items easy, as the prompt for each command includes the history ID that will represent it.

Note

You can easily see how long a series of commands took to invoke by looking at the StartExecutionTime and EndExecutionTime properties. This is a great way to get a handle on exactly how little time it took to come up with the commands that just saved you hours of manual work:

PS C:\> Get-History 65,66 | Format-Table *

Id CommandLine             StartExecutionTime    EndExecutionTime
-- -----------             ------------------    ----------------
65 dir                     10/13/2012 2:06:05 PM 10/13/2012 2:06:05 PM
66 Start-Sleep -Seconds 45 10/13/2012 2:06:15 PM 10/13/2012 2:07:00 PM

IDs provided by the Get-History cmdlet differ from the IDs given by the Windows console common history hotkeys (such as F7), because their history management techniques differ.

By default, PowerShell stores the last 4,096 entries of your command history. If you want to raise or lower this amount, set the $MaximumHistoryCount variable to the size you desire. To make this change permanent, set the variable in your PowerShell profile script.

By far, the most useful feature of PowerShell’s command history is for reviewing ad hoc experimentation and capturing it in a script that you can then use over and over. For an overview of that process (and a script that helps to automate it), see Program: Create Scripts from Your Session History.

Program: Create Scripts from Your Session History

After interactively experimenting at the command line for a while to solve a multistep task, you’ll often want to keep or share the exact steps you used to eventually solve the problem. The script smiles at you from your history buffer, but it is unfortunately surrounded by many more commands that you don’t want to keep.

Note

For an example of using the Out-GridView cmdlet to do this graphically, see Program: Interactively Filter Lists of Objects.

To solve this problem, use the Get-History cmdlet to view the recent commands that you’ve typed. Then, call Copy-History with the IDs of the commands you want to keep, as shown in Example 1-16.

Example 1-16. Copy-History.ps1

##############################################################################
##
## Copy-History
##
## From Windows PowerShell Cookbook (O'Reilly)
## by Lee Holmes (http://www.leeholmes.com/guide)
##
##############################################################################

<#

.SYNOPSIS

Copies selected commands from the history buffer into the clipboard as a script.

.EXAMPLE

PS > Copy-History
Copies the entire contents of the history buffer into the clipboard.

.EXAMPLE

PS > Copy-History -5
Copies the last five commands into the clipboard.

.EXAMPLE

PS > Copy-History 2,5,8,4
Copies commands 2,5,8, and 4.

.EXAMPLE

PS > Copy-History (1..10+5+6)
Copies commands 1 through 10, then 5, then 6, using PowerShell's array
slicing syntax.

#>

param(
    ## The range of history IDs to copy
    [int[]] $Range
)

Set-StrictMode -Version 3

$history = @()

## If they haven't specified a range, assume it's everything
if((-not $range) -or ($range.Count -eq 0))
{
    $history = @(Get-History -Count ([Int16]::MaxValue))
}
## If it's a negative number, copy only that many
elseif(($range.Count -eq 1) -and ($range[0] -lt 0))
{
    $count = [Math]::Abs($range[0])
    $history = (Get-History -Count $count)
}
## Otherwise, go through each history ID in the given range
## and add it to our history list.
else
{
    foreach($commandId in $range)
    {
        if($commandId -eq -1) { $history += Get-History -Count 1 }
        else { $history += Get-History -Id $commandId }
    }
}

## Finally, export the history to the clipboard.
$history | Foreach-Object { $_.CommandLine } | clip.exe

For more information about running scripts, see Run Programs, Scripts, and Existing Tools.

Invoke a Command from Your Session History

Problem

You want to run a command from the history of your current session.

Solution

Use the Invoke-History cmdlet (or its ihy alias) to invoke a specific command by its ID:

Invoke-History ID

To search through your history for a command containing text:

PS > #text<Tab>

To repopulate your command with the text of a previous command by its ID:

PS > #ID<Tab>

Discussion

Once you’ve had your shell open for a while, your history buffer quickly fills with useful commands. The history management hotkeys described in Customize Your Shell, Profile, and Prompt show one way to navigate your history, but this type of history navigation works only for command lines you’ve typed in that specific session. If you keep a persistent command history (as shown in Save State Between Sessions), these shortcuts do not apply.

The Invoke-History cmdlet illustrates the simplest example of working with your command history. Given a specific history ID (perhaps shown in your prompt function), calling Invoke-History with that ID will run that command again. For more information about this technique, see Customize Your Shell, Profile, and Prompt.

As part of its tab-completion support, PowerShell gives you easy access to previous commands as well. If you prefix your command with the # character, tab completion takes one of two approaches:

ID completion

If you type a number, tab completion finds the entry in your command history with that ID, and then replaces your command line with the text of that history entry. This is especially useful when you want to slightly modify a previous history entry, since Invoke-History by itself doesn’t support that.

Pattern completion

If you type anything else, tab completion searches for entries in your command history that contain that text. Under the hood, PowerShell uses the -like operator to match your command entries, so you can use all of the wildcard characters supported by that operator. For more information on searching text for patterns, see Search a String for Text or a Pattern.

PowerShell’s tab completion is largely driven by the fully customizable TabExpansion2 function. You can easily change this function to include more advanced functionality, or even just customize specific behaviors to suit your personal preferences. For more information, see Understand and Customize PowerShell’s Tab Completion.

Program: Search Formatted Output for a Pattern

While PowerShell’s built-in filtering facilities are incredibly flexible (for example, the Where-Object cmdlet), they generally operate against specific properties of the incoming object. If you are searching for text in the object’s formatted output, or don’t know which property contains the text you are looking for, simple text-based filtering is sometimes helpful.

To solve this problem, you can pipe the output into the Out-String cmdlet before passing it to the Select-String cmdlet:

Get-Service | Out-String -Stream | Select-String audio

Or, using built-in aliases:

Get-Service | oss | sls audio

In script form, Select-TextOutput (shown in Example 1-17) does exactly this, and it lets you search for a pattern in the visual representation of command output.

Example 1-17. Select-TextOutput.ps1

##############################################################################
##
## Select-TextOutput
##
## From Windows PowerShell Cookbook (O'Reilly)
## by Lee Holmes (http://www.leeholmes.com/guide)
##
##############################################################################

<#

.SYNOPSIS

Searches the textual output of a command for a pattern.

.EXAMPLE

PS > Get-Service | Select-TextOutput audio
Finds all references to "Audio" in the output of Get-Service

#>

param(
    ## The pattern to search for
    $Pattern
)

Set-StrictMode -Version 3
$input | Out-String -Stream | Select-String $pattern

For more information about running scripts, see Run Programs, Scripts, and Existing Tools.

Interactively View and Process Command Output

Problem

You want to graphically explore and analyze the output of a command.

Solution

Use the Out-GridView cmdlet to interactively explore the output of a command.

Discussion

The Out-GridView cmdlet is one of the rare PowerShell cmdlets that displays a graphical user interface. While the Where-Object and Sort-Object cmdlets are the most common way to sort and filter lists of items, the Out-GridView cmdlet is very effective at the style of repeated refinement that sometimes helps you develop complex queries. Figure 1-3 shows the Out-GridView cmdlet in action.

Out-GridView, ready to filter

Figure 1-3. Out-GridView, ready to filter

Out-GridView lets you primarily filter your command output in two ways: a quick filter expression and a criteria filter.

Quick filters are fairly simple. As you type text in the topmost “Filter” window, Out-GridView filters the list to contain only items that match that text. If you want to restrict this text filtering to specific columns, simply provide a column name before your search string and separate the two with a colon. You can provide multiple search strings, in which case Out-GridView returns only rows that match all of the required strings.

Note

Unlike most filtering cmdlets in PowerShell, the quick filters in the Out-GridView cmdlet do not support wildcards or regular expressions. For this type of advanced query, criteria-based filtering can help.

Criteria filters give fine-grained control over the filtering used by the Out-GridView cmdlet. To apply a criteria filter, click the “Add criteria” button and select a property to filter on. Out-GridView adds a row below the quick filter field and lets you pick one of several operations to apply to this property:

  • Less than or equal to

  • Greater than or equal to

  • Between

  • Equals

  • Does not equal

  • Contains

  • Does not contain

In addition to these filtering options, Out-GridView also lets you click and rearrange the header columns to sort by them.

Processing output

Once you’ve sliced and diced your command output, you can select any rows you want to keep and press Ctrl-C to copy them to the clipboard. Out-GridView copies the items to the clipboard as tab-separated data, so you can easily paste the information into a spreadsheet or other file for further processing.

In addition to supporting clipboard output, the Out-GridView cmdlet supports full-fidelity object filtering if you use its -PassThru parameter. For an example of this full-fidelity filtering, see Program: Interactively Filter Lists of Objects.

Program: Interactively View and Explore Objects

When working with unfamiliar objects in PowerShell, much of your time is spent with the Get-Member and Format-List commands—navigating through properties, reviewing members, and more.

For ad hoc investigation, a graphical interface is often useful.

To solve this problem, Example 1-18 provides an interactive tree view that you can use to explore and navigate objects. For example, to examine the structure of a script as PowerShell sees it (its abstract syntax tree):

$ps = { Get-Process -ID $pid }.Ast
Show-Object $ps

For more information about parsing and analyzing the structure of PowerShell scripts, see Parse and Interpret PowerShell Scripts.

Example 1-18. Show-Object.ps1

#############################################################################
##
## Show-Object
##
## From Windows PowerShell Cookbook (O'Reilly)
## by Lee Holmes (http://www.leeholmes.com/guide)
##
##############################################################################

<#

.SYNOPSIS

Provides a graphical interface to let you explore and navigate an object.


.EXAMPLE

PS > $ps = { Get-Process -ID $pid }.Ast
PS > Show-Object $ps

#>

param(
    ## The object to examine
    [Parameter(ValueFromPipeline = $true)]
    $InputObject
)

Set-StrictMode -Version 3

Add-Type -Assembly System.Windows.Forms

## Figure out the variable name to use when displaying the
## object navigation syntax. To do this, we look through all
## of the variables for the one with the same object identifier.
$rootVariableName = dir variable:\* -Exclude InputObject,Args |
    Where-Object {
        $_.Value -and
        ($_.Value.GetType() -eq $InputObject.GetType()) -and
        ($_.Value.GetHashCode() -eq $InputObject.GetHashCode())
}

## If we got multiple, pick the first
$rootVariableName = $rootVariableName| % Name | Select -First 1

## If we didn't find one, use a default name
if(-not $rootVariableName)
{
    $rootVariableName = "InputObject"
}

## A function to add an object to the display tree
function PopulateNode($node, $object)
{
    ## If we've been asked to add a NULL object, just return
    if(-not $object) { return }

    ## If the object is a collection, then we need to add multiple
    ## children to the node
    if([System.Management.Automation.LanguagePrimitives]::GetEnumerator($object))
    {
        ## Some very rare collections don't support indexing (i.e.: $foo[0]).
        ## In this situation, PowerShell returns the parent object back when you
        ## try to access the [0] property.
        $isOnlyEnumerable = $object.GetHashCode() -eq $object[0].GetHashCode()

        ## Go through all the items
        $count = 0
        foreach($childObjectValue in $object)
        {
            ## Create the new node to add, with the node text of the item and
            ## value, along with its type
            $newChildNode = New-Object Windows.Forms.TreeNode
            $newChildNode.Text = "$($node.Name)[$count] = $childObjectValue : " +
                $childObjectValue.GetType()

            ## Use the node name to keep track of the actual property name
            ## and syntax to access that property.
            ## If we can't use the index operator to access children, add
            ## a special tag that we'll handle specially when displaying
            ## the node names.
            if($isOnlyEnumerable)
            {
                $newChildNode.Name = "@"
            }

            $newChildNode.Name += "[$count]"
            $null = $node.Nodes.Add($newChildNode)               

            ## If this node has children or properties, add a placeholder
            ## node underneath so that the node shows a '+' sign to be
            ## expanded.
            AddPlaceholderIfRequired $newChildNode $childObjectValue

            $count++
        }
    }
    else
    {
        ## If the item was not a collection, then go through its
        ## properties
        foreach($child in $object.PSObject.Properties)
        {
            ## Figure out the value of the property, along with
            ## its type.
            $childObject = $child.Value
            $childObjectType = $null
            if($childObject)
            {
                $childObjectType = $childObject.GetType()
            }

            ## Create the new node to add, with the node text of the item and
            ## value, along with its type
            $childNode = New-Object Windows.Forms.TreeNode
            $childNode.Text = $child.Name + " = $childObject : $childObjectType"
            $childNode.Name = $child.Name
            $null = $node.Nodes.Add($childNode)

            ## If this node has children or properties, add a placeholder
            ## node underneath so that the node shows a '+' sign to be
            ## expanded.
            AddPlaceholderIfRequired $childNode $childObject
        }
    }
}

## A function to add a placeholder if required to a node.
## If there are any properties or children for this object, make a temporary
## node with the text "..." so that the node shows a '+' sign to be
## expanded.
function AddPlaceholderIfRequired($node, $object)
{
    if(-not $object) { return }

    if([System.Management.Automation.LanguagePrimitives]::GetEnumerator($object) -or
        @($object.PSObject.Properties))
    {
        $null = $node.Nodes.Add( (New-Object Windows.Forms.TreeNode "...") )
    }
}

## A function invoked when a node is selected.
function OnAfterSelect
{
    param($Sender, $TreeViewEventArgs)

    ## Determine the selected node
    $nodeSelected = $Sender.SelectedNode

    ## Walk through its parents, creating the virtual
    ## PowerShell syntax to access this property.
    $nodePath = GetPathForNode $nodeSelected

    ## Now, invoke that PowerShell syntax to retrieve
    ## the value of the property.
    $resultObject = Invoke-Expression $nodePath
    $outputPane.Text = $nodePath

    ## If we got some output, put the object's member
    ## information in the text box.
    if($resultObject)
    {
        $members = Get-Member -InputObject $resultObject | Out-String       
        $outputPane.Text += "`n" + $members
    }
}

## A function invoked when the user is about to expand a node
function OnBeforeExpand
{
    param($Sender, $TreeViewCancelEventArgs)

    ## Determine the selected node
    $selectedNode = $TreeViewCancelEventArgs.Node

    ## If it has a child node that is the placeholder, clear
    ## the placeholder node.
    if($selectedNode.FirstNode -and
        ($selectedNode.FirstNode.Text -eq "..."))
    {
        $selectedNode.Nodes.Clear()
    }
    else
    {
        return
    }

    ## Walk through its parents, creating the virtual
    ## PowerShell syntax to access this property.
    $nodePath = GetPathForNode $selectedNode 

    ## Now, invoke that PowerShell syntax to retrieve
    ## the value of the property.
    Invoke-Expression "`$resultObject = $nodePath"

    ## And populate the node with the result object.
    PopulateNode $selectedNode $resultObject
}

## A function to handle keypresses on the form.
## In this case, we capture ^C to copy the path of
## the object property that we're currently viewing.
function OnKeyPress
{
    param($Sender, $KeyPressEventArgs)

    ## [Char] 3 = Control-C
    if($KeyPressEventArgs.KeyChar -eq 3)
    {
        $KeyPressEventArgs.Handled = $true

        ## Get the object path, and set it on the clipboard
        $node = $Sender.SelectedNode
        $nodePath = GetPathForNode $node
        [System.Windows.Forms.Clipboard]::SetText($nodePath)

        $form.Close()
    }
}

## A function to walk through the parents of a node,
## creating virtual PowerShell syntax to access this property.
function GetPathForNode
{
    param($Node)

    $nodeElements = @()

    ## Go through all the parents, adding them so that
    ## $nodeElements is in order.
    while($Node)
    {
        $nodeElements = ,$Node + $nodeElements
        $Node = $Node.Parent
    }

    ## Now go through the node elements
    $nodePath = ""
    foreach($Node in $nodeElements)
    {
        $nodeName = $Node.Name

        ## If it was a node that PowerShell is able to enumerate
        ## (but not index), wrap it in the array cast operator.
        if($nodeName.StartsWith('@'))
        {
            $nodeName = $nodeName.Substring(1)
            $nodePath = "@(" + $nodePath + ")"
        }
        elseif($nodeName.StartsWith('['))
        {
            ## If it's a child index, we don't need to
            ## add the dot for property access
        }
        elseif($nodePath)
        {
            ## Otherwise, we're accessing a property. Add a dot.
            $nodePath += "."
        }

        ## Append the node name to the path
        $nodePath += $nodeName
    }

    ## And return the result
    $nodePath
}

## Create the TreeView, which will hold our object navigation
## area.
$treeView = New-Object Windows.Forms.TreeView
$treeView.Dock = "Top"
$treeView.Height = 500
$treeView.PathSeparator = "."
$treeView.Add_AfterSelect( { OnAfterSelect @args } )
$treeView.Add_BeforeExpand( { OnBeforeExpand @args } )
$treeView.Add_KeyPress( { OnKeyPress @args } )

## Create the output pane, which will hold our object
## member information.
$outputPane = New-Object System.Windows.Forms.TextBox
$outputPane.Multiline = $true
$outputPane.ScrollBars = "Vertical"
$outputPane.Font = "Consolas"
$outputPane.Dock = "Top"
$outputPane.Height = 300

## Create the root node, which represents the object
## we are trying to show.
$root = New-Object Windows.Forms.TreeNode
$root.Text = "$InputObject : " + $InputObject.GetType()
$root.Name = '$' + $rootVariableName
$root.Expand()
$null = $treeView.Nodes.Add($root)

## And populate the initial information into the tree
## view.
PopulateNode $root $InputObject

## Finally, create the main form and show it.
$form = New-Object Windows.Forms.Form
$form.Text = "Browsing " + $root.Text
$form.Width = 1000
$form.Height = 800
$form.Controls.Add($outputPane)
$form.Controls.Add($treeView)
$null = $form.ShowDialog()
$form.Dispose()

For more information about running scripts, see Run Programs, Scripts, and Existing Tools.

Store the Output of a Command into a File

Problem

You want to redirect the output of a command or pipeline into a file.

Solution

To redirect the output of a command into a file, use either the Out-File cmdlet or one of the redirection operators.

Out-File:

Get-ChildItem | Out-File unicodeFile.txt
Get-Content filename.cs | Out-File -Encoding ASCII file.txt
Get-ChildItem | Out-File -Width 120 unicodeFile.cs

Redirection operators:

Get-ChildItem > files.txt
Get-ChildItem 2> errors.txt
Get-ChildItem n> otherStreams.txt

Discussion

The Out-File cmdlet and redirection operators share a lot in common. For the most part, you can use either. The redirection operators are unique because they give the greatest amount of control over redirecting individual streams. The Out-File cmdlet is unique primarily because it lets you easily configure the formatting width and encoding.

Note

If you want to save the objects from a command into a file (rather than the text-based representation that you see on screen), see Easily Import and Export Your Structured Data.

The default formatting width and the default output encoding are two aspects of output redirection that can sometimes cause difficulty.

The default formatting width sometimes causes problems because redirecting PowerShell-formatted output into a file is designed to mimic what you see on the screen. If your screen is 80 characters wide, the file will be 80 characters wide as well. Examples of PowerShell-formatted output include directory listings (that are implicitly formatted as a table) as well as any commands that you explicitly format using one of the Format-* set of cmdlets. If this causes problems, you can customize the width of the file with the -Width parameter on the Out-File cmdlet.

The default output encoding sometimes causes unexpected results because PowerShell creates all files using the UTF-16 Unicode encoding by default. This allows PowerShell to fully support the entire range of international characters, cmdlets, and output. Although this is a great improvement on traditional shells, it may cause an unwanted surprise when running large search-and-replace operations on ASCII source code files, for example. To force PowerShell to send its output to a file in the ASCII encoding, use the -Encoding parameter on the Out-File cmdlet.

For more information about the Out-File cmdlet, type Get-Help Out-File. For a full list of supported redirection operators, see Capturing Output.

Add Information to the End of a File

Problem

You want to redirect the output of a pipeline into a file but add the information to the end of that file.

Solution

To redirect the output of a command into a file, use either the -Append parameter of the Out-File cmdlet or one of the appending redirection operators described in Capturing Output. Both support options to append text to the end of a file.

Out-File:

Get-ChildItem | Out-File -Append files.txt

Redirection operators:

Get-ChildItem >> files.txt

Discussion

The Out-File cmdlet and redirection operators share a lot in common. For the most part, you can use either. See the discussion in Store the Output of a Command into a File for a more detailed comparison of the two approaches, including reasons that you would pick one over the other.

Record a Transcript of Your Shell Session

Problem

You want to record a log or transcript of your shell session.

Solution

To record a transcript of your shell session, run the command Start-Transcript. It has an optional -Path parameter that defaults to a filename based on the current system time. By default, PowerShell places this file in the My Documents directory. To stop recording the transcript of your shell system, run the command Stop-Transcript.

Discussion

Although the Get-History cmdlet is helpful, it does not record the output produced during your PowerShell session. To accomplish that, use the Start-Transcript cmdlet. In addition to the Path parameter described previously, the Start-Transcript cmdlet also supports parameters that let you control how PowerShell interacts with the output file.

Extend Your Shell with Additional Commands

Problem

You want to use PowerShell cmdlets, providers, or script-based extensions written by a third party.

Solution

If the module is part of the standard PowerShell module path, simply run the command you want.

Invoke-NewCommand

If it is not, use the Import-Module command to import third-party commands into your PowerShell session.

To import a module from a specific directory:

Import-Module c:\path\to\module

To import a module from a specific file (module, script, or assembly):

Import-Module c:\path\to\module\file.ext

Discussion

PowerShell supports two sets of commands that enable additional cmdlets and providers: *-Module and *-PsSnapin. Snapins were the packages for extensions in version 1 of PowerShell, and are rarely used. Snapins supported only compiled extensions and had onerous installation requirements.

Version 2 of PowerShell introduced modules that support everything that snapins support (and more) without the associated installation pain. That said, PowerShell version 2 also required that you remember which modules contained which commands and manually load those modules before using them. Windows 8 and Windows Server 2012 include thousands of commands in over 50 modules—quickly making reliance on one’s memory an unsustainable approach.

PowerShell version 3 significantly improves the situation by autoloading modules for you. Internally, it maintains a mapping of command names to the module that contains them. Simply start using a command (which the Get-Command cmdlet can help you discover), and PowerShell loads the appropriate module automatically. If you wish to customize this autoloading behavior, you can use the $PSModuleAutoLoadingPreference preference variable.

When PowerShell imports a module with a given name, it searches through every directory listed in the PSModulePath environment variable, looking for the first module that contains the subdirectories that match the name you specify. Inside those directories, it looks for the module (*.psd1, *.psm1, and *.dll) with the same name and loads it.

Note

When autoloading modules, PowerShell prefers modules in the system’s module directory over those in your personal module path. This prevents user modules from accidentally overriding core functionality. If you want a module to override core functionality, you can still use the Import-Module cmdlet to load the module explicitly.

When you install a module on your own system, the most common place to put it is in the WindowsPowerShell\Modules directory in your My Documents directory. To have PowerShell look in another directory for modules, add it to your personal PSModulePath environment variable, just as you would add a Tools directory to your personal path.

For more information about managing system paths, see Modify the User or System Path.

If you want to load a module from a directory not in PSModulePath, you can provide the entire directory name and module name to the Import-Module command. For example, for a module named Test, use Import-Module c:\path\to\Test. As with loading modules by name, PowerShell looks in c:\temp\path\to for a module (*.psd1, *.psm1, or *.dll) named Test and loads it.

If you know the specific module file you want to load, you can also specify the full path to that module.

If you want to find additional commands, there are several useful resources available.

PowerShell Community Extensions

Located here, the PowerShell Community Extensions project contains a curated collection of useful and powerful commands. It has been written by a handful of volunteers, many of them Microsoft MVPs.

The Technet Script Center Gallery

Located here, the TechNet Script Center Gallery offers a well-indexed and well-organized collection of PowerShell scripts.

PoshCode

Located here, PoshCode contains thousands of scripts and script snippets—of both high and low quality.

Use Commands from Customized Shells

Problem

You want to use the commands from a PowerShell-based product that launches a customized version of the PowerShell console, but in a regular PowerShell session.

Solution

Launch the customized version of the PowerShell console, and then use the Get-Module and Get-PsSnapin commands to see what additional modules and/or snapins it loaded.

Discussion

As described in Extend Your Shell with Additional Commands, PowerShell modules and snapins are the two ways that third parties can distribute and add additional PowerShell commands. Products that provide customized versions of the PowerShell console do this by calling PowerShell.exe with one of three parameters:

  • -PSConsoleFile, to load a console file that provides a list of snapins to load.

  • -Command, to specify an initial startup command (that then loads a snapin or module)

  • -File, to specify an initial startup script (that then loads a snapin or module)

Regardless of which one it used, you can examine the resulting set of loaded extensions to see which ones you can import into your other PowerShell sessions.

Detecting loaded snapins

The Get-PsSnapin command returns all snapins loaded in the current session. It always returns the set of core PowerShell snapins, but it will also return any additional snapins loaded by the customized environment. For example, if the name of a snapin you recognize is Product.Feature.Commands, you can load that into future PowerShell sessions by typing Add-PsSnapin Product.Feature.Commands. To automate this, add the command into your PowerShell profile.

If you are uncertain of which snapin to load, you can also use the Get-Command command to discover which snapin defines a specific command:

PS > Get-Command Get-Counter | Select PsSnapin

PSSnapIn
--------
Microsoft.PowerShell.Diagnostics

Detecting loaded modules

Like the Get-PsSnapin command, the Get-Module command returns all modules loaded in the current session. It returns any modules you’ve added so far into that session, but it will also return any additional modules loaded by the customized environment. For example, if the name of a module you recognize is ProductModule, you can load that into future PowerShell sessions by typing Import-Module ProductModule. To automate this, add the command into your PowerShell profile.

If you are uncertain of which module to load, you can also use the Get-Command command to discover which module defines a specific command:

PS > Get-Command Start-BitsTransfer | Select Module

Module
------
BitsTransfer

Save State Between Sessions

Problem

You want to save state or history between PowerShell sessions.

Solution

Subscribe to the PowerShell.Exiting engine event to have PowerShell invoke a script or script block that saves any state you need.

To have PowerShell save your command history, place a call to Enable-HistoryPersistence in your profile, as in Example 1-19.

Example 1-19. Enable-HistoryPersistence.ps1

##############################################################################
##
## Enable-HistoryPersistence
##
## From Windows PowerShell Cookbook (O'Reilly)
## by Lee Holmes (http://www.leeholmes.com/guide)
##
##############################################################################

<#

.SYNOPSIS

Reloads any previously saved command history, and registers for the
PowerShell.Exiting engine event to save new history when the shell
exits.

#>

Set-StrictMode -Version 3

## Load our previous history
$GLOBAL:maximumHistoryCount = 32767
$historyFile = (Join-Path (Split-Path $profile) "commandHistory.clixml")
if(Test-Path $historyFile)
{
    Import-CliXml $historyFile | Add-History
}

## Register for the engine shutdown event
$null = Register-EngineEvent -SourceIdentifier `
    ([System.Management.Automation.PsEngineEvent]::Exiting) -Action {

    ## Save our history
    $historyFile = (Join-Path (Split-Path $profile) "commandHistory.clixml")
    $maximumHistoryCount = 1kb

    ## Get the previous history items
    $oldEntries = @()
    if(Test-Path $historyFile)
    {
        $oldEntries = Import-CliXml $historyFile -ErrorAction SilentlyContinue
    }

    ## And merge them with our changes
    $currentEntries = Get-History -Count $maximumHistoryCount
    $additions = Compare-Object $oldEntries $currentEntries `
        -Property CommandLine | Where-Object { $_.SideIndicator -eq "=>" } |
        Foreach-Object { $_.CommandLine }

    $newEntries = $currentEntries | ? { $additions -contains $_.CommandLine }

    ## Keep only unique command lines. First sort by CommandLine in
    ## descending order (so that we keep the newest entries,) and then
    ## re-sort by StartExecutionTime.
    $history = @($oldEntries + $newEntries) |
        Sort -Unique -Descending CommandLine | Sort StartExecutionTime

    ## Finally, keep the last 100
    Remove-Item $historyFile
    $history | Select -Last 100 | Export-CliXml $historyFile
}

Discussion

PowerShell provides easy script-based access to a broad variety of system, engine, and other events. You can register for notification of these events and even automatically process any of those events. In this example, we subscribe to the only one currently available, which is called PowerShell.Exiting. PowerShell generates this event when you close a session.

This script could do anything, but in this example we have it save our command history and restore it when we launch PowerShell. Why would we want to do this? Well, with a rich history buffer, we can more easily find and reuse commands we’ve previously run. For two examples of doing this, see Examples and .

Example 1-19 takes two main actions. First, we load our stored command history (if any exists). Then, we register an automatic action to be processed whenever the engine generates its PowerShell.Exiting event. The action itself is relatively straightforward, although exporting our new history does take a little finesse. If you have several sessions open at the same time, each will update the saved history file when it exits. Since we don’t want to overwrite the history saved by the other shells, we first reload the history from disk and combine it with the history from the current shell.

Once we have the combined list of command lines, we sort them and pick out the unique ones before storing them back in the file.

For more information about working with PowerShell engine events, see Create and Respond to Custom Events.

Get Windows PowerShell Cookbook, 3rd Edition now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.