This question already has an answer here:
I want to be able to tell what path my executing script was run from.
This will often not be $pwd.
I need to call other scripts that are in a folder structure relative to my script and while I could hard code the paths, that's both distasteful and a bit of a pain in the neck when trying to promote from "dev" to "test" to "production".
Thank you msorens! This really helped me with my custom module. In case anyone is interested in making their own, here is how mine is structured.
You then reference MyScriptFile.ps1 in MyModule.psd1. Referencing the .ps1 in the NestedModules array will place the functions in the module session state rather than the global session state. (How to Write a Module Manifest)
Content of MyScriptFile.ps1
The try/catch hides the error from Export-ModuleMember when running MyScriptFile.ps1
Copy the MyModule directory to one of the paths found here $env:PSModulePath
We've been using code like this in most of our scripts for several years with no problems:
This is one of those oddities (to my mind at least) in PS. I'm sure there is a perfectly good reason for it, but it still seems odd to me. So:
If you are in a script but not in a function then $myInvocation.InvocationName will give you the full path including the script name. If you are in a script and inside a function then $myInvocation.ScriptName will give you the same thing.
You tagged your question for Powershell version 1.0, however, if you have access to Powershell version 3.0 you know have
$PSCommandPath
and$PSScriptRoot
which makes getting the script path a little easier. Please refer to the "OTHER SCRIPT FEATURES" section on this page for more information.I ran into the same issue recently. The following article helped me solve the problem: http://blogs.msdn.com/powershell/archive/2007/06/19/get-scriptdirectory.aspx
If you're not interested in how it works, here's all the code you need per the article:
And then you get the path by simply doing:
The ubiquitous script originally posted by Jeffrey Snover of the PowerShell team (given in Skyler's answer) and the variations posted by Keith Cedirc, and EBGreen, all suffer from a serious drawback--whether the code reports what you expect depends on where you call it!
My code below overcomes this problem by simply referencing script scope instead of parent scope:
To illustrate the problem, I created a test vehicle that evalutes the target expression in four different ways. (The bracketed terms are the keys to the following result table.)
The last two columns show the result of using script scope (i.e. $script:) or with parent scope (with -scope 1). A result of "script" means that the invocation correctly reported the location of the script. The "module" result means the invocation reported the location of the module containing the function rather than the script that called the function; this indicates a drawback of both functions that you cannot put the function in a module.
Setting the module issue aside the remarkable observation from the table is that using the parent scope approach fails most of the time (in fact, twice as often as it succeeds).
Finally, here is the test vehicle:
Contents of ScriptDirFinder.ps1:
Contents of ScriptDirFinder.psm1:
I am not familiar with what was introduced in PowerShell 2, but it could very well be that script scope did not exist in PowerShell 1, at the time Jeffrey Snover published his example.
I was surprised when, though I found his code example proliferated far and wide on the web, it failed immediately when I tried it! But that was because I used it differently than Snover's example (I called it not at script-top but from inside another function (my "nested twice" example).)
2011.09.12 Update
You can read about this with other tips and tricks on modules in my just-published article on Simple-Talk.com: Further Down the Rabbit Hole: PowerShell Modules and Encapsulation.