Archives

Archives / 2013
  • SPListCollection ContainsList Extension Method

    This is such a simple thing but something every SharePoint developer should have in their toolkit (well, actually, this is something Microsoft should put into the product). The SPFieldCollection has a nice method called ContainsField so you can check for the existence of a field (without throwing an exception if you get the spelling wrong). This is my version of the same method for SharePoint Lists.

     public static class SPListExtensions
    {
    public static bool ContainsList(this SPListCollection lists, string displayName)
    {
    try
    {
    return lists.Cast<SPList>().Any(list => string.Equals(list.Title, displayName, StringComparison.OrdinalIgnoreCase);
    }
    catch (Exception)
    {
    return false;
    }
    }
    }
    Enjoy.

  • Enhancing your SharePoint Team Site Homepage

    A team site can be a boring place. Just a site with some documents, a list or two, maybe a calendar. Here’s a super simple way to make the page a little more interesting looking.

    Your team site might look like this right now, with a calendar on the home page:

    before

    Simple and effective but bland. Also the title just blends into the background doesn’t it?

    Edit the web part and under Advanced settings, set the Title Icon Image URL to the same value as the Calendar Icon Image URL:

    settings

    Now the calendar shows a small icon next to the title and breaks up the page a little.

    after

    Simple but easy to do. By default, the Title icon is blank but IMHO it should be defaulted to the type of list that’s being shown (Calendar, Task List, Document Library, etc.). Like I said, it’s simple but it breaks up the page and also is a visual reminder of what you’re looking at. It also shows where the clickable title is for a list or library that’s been put on a page. When you have a page full of web parts, it’s not immediately obvious where the title is in a sea of text. A little graphic goes a long way for readability.

    You can use any icon for any list but I like using the one suited to the task. Here’s a list of the icons to use for each list/library type:

    List/Library Type Icon
    Custom List /_layouts/images/itgen.gif
    Calendar /_layouts/images/itevent.png
    Contact List /_layouts/images/itcontct.gif
    Site /_layouts/images/SharePointFoundation16.png
    Document Library /_layouts/images/itdl.png
    Announcements /_layouts/images/itann.png
    Discussion Board /_layouts/images/itann.png
    Issue Tracker /_layouts/images/itissue.png
    Picture Library /_layouts/images/itil.png
    Links /_layouts/images/itlink.png
    Tasks /_layouts/images/ittask.png
    Survey /_layouts/images/itsurvey.png

    Enjoy!

  • Virtual Machine Tips for running SharePoint

    Just a few simple Virtual Machine tips when developing in SharePoint

    • My VMs are 40GB in size and pre-allocated. After installing SharePoint, SQL Server, and Visual Studio and all the tools I need I have about 10GB for data which is more than sufficient for most jobs. The pre-allocation helps performance but you need a lot of disk space so keep that in mind when planning your VM.
    • I run my VMs disconnected with their own Domain Controller (for development I usually just run a single VM with DC, SQL, and SharePoint all in one). I reconnect them via a bridge to the host network adapter when I'm at home off my corporate network. This is to get patches and keep things up to date. Running disconnected is nice because you know when things fail and why. Got a rogue web part that's behaving strangely? Put it in your VM and take a look at it to make sure it's not relying on some external service (which may be throwing proxy errors or something from your corporate network)
    • I rely on source control to store my code and only pull down copies of current projects when I'm ready to work on something and keep things fresh with mulitple check-ins per day. With VMWare the nice thing is that I can mount the VM as a disk image in case I need to pull something off (without having to fire up the VM). Having a source control system like GIT or Mercurial instead of Subversion or VSS will let you do local check-ins then you can sync when you're done at the end of the day.
    • VMs on the local drive (if you're running a SSD) are waaaaaaaaaaay faster than on an external drive. If you have to use an external drive for your VMs (I've since stopped doing this with the increased size of SSDs these days) then go with eSATA for the best performance (although some people will argue USB 3.0 is faster). Pre-allocating the VM on an external drive will help with performance as well.
    • Give your VM lots of memory. Any SharePoint developer must be running at *least* 8GB on their host machine. Give your VM 4-5 GB. No watching movies while you work, defer that to another machine. For SharePoint 2013 you really need to run 16GB on your host machine and give your VMs 8-12 GB of RAM.
    • Copying large files into a VM (like an installer) then deleting it will cause fragmentation that you might not get back during regular usage. Make sure you use your virtualization tools to defragment your VMs on a regular basis.
    • A really great tool to keep your VMs under control size wise is SpaceSniffer. It visually shows you where things are gobbling up space in the OS so you can pinpoint things that you don't need and zap 'em! Get it. It's free!

    That's it for now. Just a few simple tips that might help out. Happy developing!

  • ASP.NET Training Videos Delivered via SharePoint

    Recently Scott Hanselman posted on his blog an ASP.NET Jump Start session they had which featured 9 videos (over 8 hours of content) on developing apps with ASP.NET. This is a great resource and I wanted to share it with the rest of my team. Problem is that a) the team generally doesn’t have access to video content on the web as it’s generally blocked by proxy b) streaming an hour video over the Internet might be okay for one but not for an entire team and c) there must be a better way to share this other than passing out links to Scott’s blog or the Channel 9 site.

    We have a team site in SharePoint so I thought this would be a great opportunity to share the information that way. The problem was that even at medium resolution the files are just too big to host inside of SharePoint.

    Sidebar. There have been a lot of discussions about “Large file support” in SharePoint 2010 (and beyond) and that’s great but for me the bottom line is that extremely large files (over the default 50MB in size) just isn’t meant for an Intranet. I can hear the arguments now about it but a few things to consider with large files (specifically media files like videos). Uploading a 100MB file to the server means 100MB of memory gets gobbled up by the w3wp.exe process (the process attached to the Application Pool running your site) during the entire time the file is being uploaded. LANs are fast but even a 100MB file only gets uploaded at a certain speed (regardless of how big your pipe is). Also uploading the video commits the user to that web front end (usually pegging the server or process) so your load balancing is shot. In short, many people bump up the maximum size of SharePoint’s default of 50MB without taking any of these things into consideration then wonder why their amazing Intranet site is running slow (and usually they toss more memory on the front-ends thinking it will help). Basically 50MB should be the limit for files on any web application when users are uploading.

    In any case, an option is to specify the files via UNC paths which means that a) I can just use the Windows Desktop top copy the files so no size limitations and b) I can address the files via the file:// protocol. This works much for most Intranets and large files and should be considered over stuffing a file into SharePoint (regardless of how you get it in). Remember that a 100MB file in SharePoint generally takes up at least 150MB of space (and if the file is versioned, eek! just watch your content database explode in size!).

    So my first step was to download the files and put them onto a file share. Simple enough. 9 files. I grabbed the medium size files but you can get the HD ones or whatever you want (I just didn’t want to wait around for 500MB per file to download with the high quality WMV files). Medium quality is good enough for training as you can clearly see the code in the IDE and not be annoyed by pixel artifacts on the playback.

    Now that I had my 9 videos of content on the corporate LAN it was time to set something up to display them. You can get way fancier than I did here with some jQuery, doing some Client Object Model code to write out a video carousel but this is a simple solution that required no code (and I do mean absolutely zero code) and a few minutes of time. In fact it took longer to copy the files than it did to setup the list.

    First create a Custom List on the site you want to server your video catalog up from. Then I added three fields (in addition to the default Title field):

    1. Video. This was set to Hyperlink field that would be the UNC path to the video itself on the file share.
    2. Thumbnail. This was a Hyperlink field formatted as a picture and would offer up a snapshot of what the user was going to see.
    3. Description. A multiline field set to Text only that would hold a brief description of the video.

    Here’s the Edit form for a video item:

    • The title is whatever you want (it’s not used for the catalog)
    • The Video is a hyperlink to the file on the file share. This will take the form of “file://” instead of “http://” and point to your UNC path to the file. Put the full title (or whatever the user is going to see to click on) in the description field. With Hyperlink fields if you leave the description field blank it gets automatically filled in with the full link (which isn’t very user friendly).
    • The Thumbnail is a hyperlink to an image. You can choose to make your own (if your content is your own). I cheated here and just grabbed the image directly off of Scott’s blog (which is up on some Microsoft content delivery network location). As this field is formatted as a Picture it’ll just display the image so make sure it points to an image file SharePoint recognizes (JPEG is probably preferred here).
    • The Description field is just copied straight off the blog. Again, change this to whatever you want so users know what the video is about.

    Once you populate the list with the videos you almost ready to go. Here’s the default view of the list:

    This is okay but not very user friendly for viewing videos. Users might click on the Title field which would open up the List Item instead of launching the video. So instead either create a new view or modify the default one. If you create a new view you might want to set it as a new default so when a user visits the list they see the right view.

    In your new view we’ll set a few options:

    • Set the first three columns to be Thumbnail, Video, and Description (in this order). This is to create the catalog view of the world so you might want another view for editing content (in case you’re adding something or want to update a value).
    • Turn off Tabular View as we don’t need it here
    • Under Style choose “Boxed, no labels”

    Now here’s the updated view in the browser:

    Pretty slick and only took 5 minutes to build the view. Users click on the title and the video launches in Windows Media Player (or whatever player is associated with your video file format you’re pointing to).

    Note that this posts talks about a specific set of files for a solution but the video content is up to you. If you have some high quality/large format audio or video files in your organization and want to surface them up in a catalog this solution might work for you. It’s not about the content, it’s about serving up that content.

    Enjoy!

  • Convention over Configuration with MVC and Autofac

    One of the key things to wrap your head around when doing good software development using frameworks like ASP.NET MVC is the idea of convention over configuration (or coding by convention).

    The idea is that rather than messing around with configuration files about where to find things, how to register IoC containers, etc. that we used to do, you follow a convention, a way of doing things. The example normally expressed is a class in your model called “Sale” and a database named “Sales”. If you don’t do anything that’s how the system will work but if you decide to change the name of the database to “Products_Sold” then you need some kind of configuration to tell the system how to find the backend database. Otherwise it can naturally find it based on the naming strategy of your domain.

    MVC does this by default. When you create a controller named “Home” the class is HomeController and the views of the HomeController are found in /Views/Home for your project. When dealing with IoC containers there’s the task of registering your types to resolve correct. So let’s take a step back and take a look at a project with several repositories in it (a repository here being some kind of abstraction over your data store).

    Here we have under our Models: Customer, Invoice, and Product with their respective classes, repositories, and interfaces:

    image

    This might be how your project is typically setup. I want to inject the dependencies on my repositories into my controller (via my IoC system) like this:

       1:  public class HomeController : Controller 
       2:  { 
       3:      private readonly ICustomerRepository _customerRepository; 
       4:      private readonly IInvoiceRepository _invoiceRepository; 
       5:      private readonly IProductRepository _productRepository;
       6:   
       7:      public HomeController(
       8:  ICustomerRepository customerRepository,
       9:  IInvoiceRepository invoiceRepository, 
      10:  IProductRepository productRepository) 
      11:      { 
      12:          _customerRepository = customerRepository; 
      13:          _invoiceRepository = invoiceRepository; 
      14:          _productRepository = productRepository; 
      15:      }
      16:   
      17:      // 
      18:      // GET: /Home/ 
      19:      public ActionResult Index() 
      20:      { 
      21:          return View(); 
      22:      } 
      23:  }

    Then somewhere in my controller I’ll use the various repositories to fetch information and present it to the user (or write back values gathered from the user). How do my IoC know how to resolve an ICustomerRepository object?

    Here’s how I have my IoC engine setup for this sample using Autofac. You can use any IoC engine you want but I find Autofac works well with MVC. First in Global.asax.cs I just follow the same pattern that the default projects setup and add a new static class called ContainerConfig.RegisterContainer()

       1:  public class MvcApplication : System.Web.HttpApplication 
       2:  { 
       3:      protected void Application_Start() 
       4:      { 
       5:          AreaRegistration.RegisterAllAreas();
       6:   
       7:          WebApiConfig.Register(GlobalConfiguration.Configuration); 
       8:          FilterConfig.RegisterGlobalFilters(GlobalFilters.Filters); 
       9:          RouteConfig.RegisterRoutes(RouteTable.Routes); 
      10:          BundleConfig.RegisterBundles(BundleTable.Bundles); 
      11:          ContainerConfig.RegisterContainer(); 
      12:      } 
      13:  }

    Next is setting up Autofac. First add the Autofac MVC4 Integration package either through the NuGet UI or the Package Manager Console:

    PM> Install-Package Autofac.Mvc4

    Next here’s my new ContainerConfig class I created which will register all the types I need:

       1:  public class ContainerConfig 
       2:  { 
       3:      public static void RegisterContainer() 
       4:      { 
       5:          var builder = new ContainerBuilder(); 
       6:          builder.RegisterControllers(Assembly.GetExecutingAssembly()); 
       7:          builder.RegisterType<CustomerRepository>().As<ICustomerRepository>(); 
       8:          builder.RegisterType<InvoiceRepository>().As<IInvoiceRepository>(); 
       9:          builder.RegisterType<ProductRepository>().As<IProductRepository>(); 
      10:          var container = builder.Build(); 
      11:          DependencyResolver.SetResolver(new AutofacDependencyResolver(container)); 
      12:      }
      13:   
      14:  }

    What’s going on here:

    Line 5 Create a new ContainerBuilder
    Line 6 Register all the controllers using the assembly object
    Line 7-9 Register each repository
    Line 10 Build the container
    Line 11 Set the default resolver to use Autofac

    Pretty straight forward but here are the issues with this approach:

    1. I have to keep going back to my ContainerConfig class adding new repositories as the system evolves. This means not only do I have to add the classes/interfaces to my system, I also have to remember to do this configuration step. New developers on the project might not remember this and the system will blow up when it can’t figure out how to resolve INewRepository
    2. I have to pull in a new namespace (assuming I follow the practice of namespace = folder structure) into the ContainerConfig class and whatever controller I add the new repository to.
    3. Repositories are scattered all over my solution (and in a big solution this can get a little ugly)

    A little messy. We can do better with convention over configuration.

    First step is to move all of your repositories into a new folder called Repositories. With ReSharper you can just use F6 to do this and it’ll automatically move and fix the namespaces for you, otherwise use whatever add-in your want or move it manually. This includes both the classes and interfaces. Here’s what our solution looks like after the move:

    image

    Pretty simple here but how does this affect our code? Really minimal. In the controller(s) that you’re injecting the repository into, you just have to remove all the old namespaces and replace it with one (whatever namespace your repositories live in).

    The other change is how you register your ContainerConfig. Here’s the updated version:

       1:  public class ContainerConfig 
       2:  { 
       3:      public static void RegisterContainer() 
       4:      { 
       5:          var builder = new ContainerBuilder(); 
       6:          builder.RegisterControllers(Assembly.GetExecutingAssembly()); 
       7:          builder.RegisterAssemblyTypes(Assembly.GetExecutingAssembly())
       8:              .Where(x => x.Namespace.EndsWith(".Repositories"))
       9:              .AsImplementedInterfaces(); 
      10:          var container = builder.Build(); 
      11:          DependencyResolver.SetResolver(new AutofacDependencyResolver(container)); 
      12:      } 
      13:  }

    Note that a) we only have one line to register all the repositories now and b) the namespace dependency we had in our file is now gone.

    The call to RegisterAssemblyTypes above using the convention of looking for any class/interface in a namespace that ends with “.Repositories” and then simply registers them all.

    So for the new developer on the project the instructions to them are to just create new repository classes and interfaces in the Repositories folder. That’s it. No configuration, no mess.

    Hope that helps!

  • Defaulting Values in a Multi-Lookup Form in SharePoint

    This was a question asked on the MSDN Forums but I thought it was worthy of a blog post as I could get more in depth with the explanation and show some pretty pictures (plus the fact I’ve never done it so thought it would be fun).

    The problem was a user wanted to default multiple values in a lookup field in SharePoint. First problem, there are no defaults in a lookup field. Second problem, how do you do default multiple values?

    First we’ll start with the setup. Create yourself a list which will hold the lookup values. In this case it’s a list of country names but it can be anything you want. Just a custom list with the Title field is enough.

    image

    Now we need a list with a lookup column to select our countries from. Create another custom list and add a column to it that looks something like this. Here’s the name and type:

    image

    And here’s the additional column settings where we get our information from (MultiLookupDefaultSpikeSource is the name of the list we created to hold our values)

    image

    Here’s what our form looks like when we add a new item:

    image

    Thinking about the problem I first though we could manipulate the form in SharePoint Designer but realized that the Form Web Part is going to retrieve all of our values from the list, defaults, etc. and really what we need to do is manipulate the list at runtime in the DOM.

    It’s jQuery to the RESCUE!

    First we take a look at the original state of the form to find our list boxes. Here’s the snippet we’re interested in, the first listbox:

     <select   
     name="ctl00$m$g_478fe6d2_8fdb_48e8_be57_7739de1c3b8f$ctl00$ctl05$ctl01$ctl00$ctl00$ctl04$ctl00$ctl00$SelectCandidate"   
     title="Country possible values"   
     id="ctl00_m_g_478fe6d2_8fdb_48e8_be57_7739de1c3b8f_ctl00_ctl05_ctl01_ctl00_ctl00_ctl04_ctl00_ctl00_SelectCandidate"   
     style="width: 143px; height: 125px; overflow: scroll;"   
     ondblclick="GipAddSelectedItems(ctl00_m_g_478fe6d2_8fdb_48e8_be57_7739de1c3b8f_ctl00_ctl05_ctl01_ctl00_ctl00_ctl04_ctl00_ctl00_MultiLookupPicker_m); return false"   
     onchange="GipSelectCandidateItems(ctl00_m_g_478fe6d2_8fdb_48e8_be57_7739de1c3b8f_ctl00_ctl05_ctl01_ctl00_ctl00_ctl04_ctl00_ctl00_MultiLookupPicker_m);"   
     multiple="multiple">   
     <OPTION title=Africa selected value=5>Africa</OPTION>   
     <OPTION title=Asia value=1>Asia</OPTION>   
     <OPTION title=Europe value=3>Europe</OPTION>   
     <OPTION title=India value=4>India</OPTION>   
     <OPTION title=Ireland value=6>Ireland</OPTION>   
     <OPTION title=Singapore value=2>Singapore</OPTION>   
     </select>  
    

    We can see that it has an ID that ends in “_SelectCandidate” so we’ll use this for selection.

    Another part of the puzzle is a hidden set of fields that store the actual values used in the list. There are three of them and they’re well documented in a blog post here by Marc Anderson on SharePoint Magazine. In it he talks about multiselect columns and breaks down the three hidden fields used (the current set of values, the complete set of values, and the default values).

    The second listbox looks like this:

     <select   
     name="ctl00$m$g_478fe6d2_8fdb_48e8_be57_7739de1c3b8f$ctl00$ctl05$ctl01$ctl00$ctl00$ctl04$ctl00$ctl00$SelectResult"   
     title="Country selected values"   
     id="ctl00_m_g_478fe6d2_8fdb_48e8_be57_7739de1c3b8f_ctl00_ctl05_ctl01_ctl00_ctl00_ctl04_ctl00_ctl00_SelectResult"   
     style="width: 143px; height: 125px; overflow: scroll;"   
     ondblclick="GipRemoveSelectedItems(ctl00_m_g_478fe6d2_8fdb_48e8_be57_7739de1c3b8f_ctl00_ctl05_ctl01_ctl00_ctl00_ctl04_ctl00_ctl00_MultiLookupPicker_m); return false" \   
     onchange="GipSelectResultItems(ctl00_m_g_478fe6d2_8fdb_48e8_be57_7739de1c3b8f_ctl00_ctl05_ctl01_ctl00_ctl00_ctl04_ctl00_ctl00_MultiLookupPicker_m);"   
     multiple="multiple">  
    

    Easy enough. It has an ID that contains “_SelectResult”.

    Now a quick jQuery primer when selecting items:

    • $("[id='foo']"); // id equals 'foo'
    • $("[id!='foo']") // id does not equal 'foo'
    • $("[id^='foo']") // id starts with 'foo'
    • $("[id$='foo']") // id ends with 'foo'
    • $("[id*='foo']") // id contains 'foo'

    Simple. We want to find the control that ends with “_SelectCandidate” and remove some items, then find the control that ends with “_SelectResult” and append our selected items.

    So a few lines of heavily commented JavaScript:

     $(document).ready(function(){  
       // define the items to add to the results (i.e already selected) this the visual part only   
       var $resultOptions = "<OPTION title=Africa value=5>Africa</OPTION><OPTION title=India value=4>India</OPTION><OPTION title=Ireland value=6>Ireland</OPTION>";   
       // this is the list of initial items (matching the ones above) that are used when the item is saved   
       var $resultSpOptions = "5|tAfrica|t4|tIndia|t6|tIreland";   
       // find the possible values control   
       var possibleValues = $("[id$='_SelectCandidate']");  
       // remove 1st option (Africa)   
       $("[id$='_SelectCandidate'] option:eq(0)").remove();  
       // remove 3rd option (India)   
       $("[id$='_SelectCandidate'] option:eq(2)").remove();  
       // remove 3rd option (Ireland)   
       $("[id$='_SelectCandidate'] option:eq(2)").remove();  
       // set selected value to asia (value 1)   
       possibleValues.val(1)  
       // append the new options to our results (this updates the display only of the second list box)   
       $("[id$='_SelectResult']").append($resultOptions);  
       // append the new options to our hidden field (this sets the values into the list item when saving)   
       $("[id$='MultiLookupPicker']").val($resultSpOptions);   
     });  
    

    SharePoint 2010 supports editing NewForm.aspx (and the other out-of-the-box forms) in the browser. One option is to modify the list and under advanced settings you can disable “Launch forms in a dialog”. This will launch the form like a regular web page. However that’s 3 or 4 steps and you have to go back and change it when you’re done.

    Instead just visit the new form directly:

    http://sitename/listname/NewForm.aspx

    From this page select Site Actions | Edit Page. Now you can add a Content Editor Web Part to the page. When adding JavaScript I point the Content Link to the .js file (that I upload somewhere like Style Library or the Assets library if you have one) rather than trying to put JavaScript into the Content Editor Web Part. This way a) I can edit the JavaScript outside of the page by loading it up in SharePoint Designer or even upload a new .js file to the library and b) I can debug the JavaScript independently of the NewForm.aspx page (or whatever page I’m adding the .js file to)

    The result:

    image

    When you save the record, the three default options are saved as well (this was set by the JavaScript).

    Hope that helps!

  • Where tips in LINQ

    These might be old but as I was going through doing some code reviews and optimizations I thought I would share with the rest of the class.

    Count() > 0 vs. Any

    This is a bit of heated debate but as you dive in the LINQ world you'll start seeing simpler ways to write things. LINQ itself gets you away from writing loops for example (sometimes). One thing I notice in code are things like this:

    if(entity.Where(some condition).Count() > 0)

    When you could just write this instead:

    if(entity.Any(some condition))

    For me, writing *less* code is *more* important. It's a cornerstone in refactoring. Making your code more readable, more succinct. If you can scan code quicker while troubleshooting a problem or trying to figure out where to add an enhancement, all the better.

    Donn Felker wrote  about the Count() vs. Any() discussion here back in 2009 and it was discussed on StackOverflow here in case you're wondering if Any() is more performant than Count() (it generally is). So my advice is use Any() to make your code potentially easier to read unless there's a performance problem.

    Another simple tip for code reduction is this one.

    var x = entity.Where(some condition).FirstOrDefault();

    Becomes:

    var x = entity.FirstOrDefault(some condition);

    I know it's simple but again it's about readability. The better you can scan someone elses new code you've never seen before is less time stumbling over the syntax and more about what the intention is.

    Yes, these tips are old but you would be surprised seeing new developers write new code with them and wonder why?

  • Windows 8 and the Lethbridge Technology User Group

    Join me and 83,517 screaming nerds (everyone in the city is attending and a geek right?) on Thursday February 21st from 3-5pm to talk about building Metro style apps for WIndows 8. Here's what we'll be covering.

    Windows 8 Platform for Metro Style Apps

    Windows 8 is Windows re-imagined. Join this session to learn about the new platform for building Metro style apps. Get an understanding of the platform design tenets, the programming language choices and the integration points with the operating system and across Metro style apps

    Everything Web Developers Must know to build Metro Style Apps

    Learn how you can use your web skills to build Windows 8 Metro style apps using HTML5, CSS3 and JavaScript. In this session you’ll discover how to harness the rich capabilities of Windows 8 through JavaScript and the Windows Runtime. You will learn about navigation, user experience patterns and controls, inherent async design and the seamless integration with the operating system that will let you create great Metro style apps.

    I promise fun times, chaos monkeys, and live kitten juggling as per my usual presentations.

  • PowerShell Tools - Removing Orphaned Users from SharePoint

    Here’s a script that will walk through all Site Collections in all Web Applications (i.e. your entire farm) and delete any user from the Site Collection that isn’t in ActiveDirectory anymore.

    Note this will not remove them from their user profiles, it just cleans up Site Collections. If you want some great info on how the User Profile service works and the My Site Cleanup Job then check out these resources:

    Here’s the code:

    [int]$GLOBAL:TotalUsersUpdated = 0;
     
    function Check_User_In_ActiveDirectory([string]$LoginName, [string]$domaincnx)
    {
        $returnValue = $false
        $strFilter = "(&(|(objectCategory=user)(objectCategory=group))(samAccountName=$LoginName))"
        $objDomain = New-Object System.DirectoryServices.DirectoryEntry($domaincnx)
     
        $objSearcher = New-Object System.DirectoryServices.DirectorySearcher
        $objSearcher.SearchRoot = $objDomain
        $objSearcher.PageSize = 1000
        $objSearcher.Filter = $strFilter
        $objSearcher.SearchScope = "Subtree"
     
        $colResults = $objSearcher.FindAll()
     
        if($colResults.Count -gt 0)
        {
            $returnValue = $true
        }
     
        return $returnValue
    }
     
    function ListOrphanedUsers([string]$SiteCollectionURL, [string]$mydomaincnx)
    {
        [System.Reflection.Assembly]::LoadWithPartialName("Microsoft.SharePoint") > $null
        $site = new-object Microsoft.SharePoint.SPSite($SiteCollectionURL)
        $web = $site.openweb()
     
        Write-Host "SiteCollectionURL:", $SiteCollectionURL
     
        $siteCollUsers = $web.SiteUsers
        Write-host "SiteUsers:", $siteCollUsers.Count
        
        #Create array to hold non-existant users
        $usersToRemove = @()
     
        foreach($MyUser in $siteCollUsers)
        {
            if(($MyUser.LoginName.ToLower() -ne "sharepoint\system") -and 
                ($MyUser.LoginName.ToLower() -ne "nt authority\authenticated users") -and 
                ($MyUser.LoginName.ToLower() -ne "nt authority\local service"))
            {
                $UserName = $MyUser.LoginName.ToLower()
                $Tablename = $UserName.split("\")
                
                $returncheck = Check_User_In_ActiveDirectory $Tablename[1] $mydomaincnx 
                if($returncheck -eq $False)
                {
                    Write-Host "User does not exist", $MyUser.LoginName, "on domain"
                    $usersToRemove = $usersToRemove + $MyUser.LoginName
                    $GLOBAL:TotalUsersUpdated += 1;
                }
            }
        }
        
        foreach($u in $usersToRemove)
        {
            Write-Host "Removing", $u, "from site collection", $SiteCollectionURL
            $siteCollUsers.Remove($u)
        }
     
        $web.Dispose()
        $site.Dispose()
    }
     
    function ListOrphanedUsersForAllColl([string]$WebAppURL, [string]$DomainCNX)
    {
        [System.Reflection.Assembly]::LoadWithPartialName("Microsoft.SharePoint") > $null
     
        $Thesite = new-object Microsoft.SharePoint.SPSite($WebAppURL)
        $oApp = $Thesite.WebApplication
        Write-host "Total Site Collections:", $oApp.Sites.Count
     
        $i = 0
        foreach ($Sites in $oApp.Sites)
        {
            $i = $i + 1
            Write-Host "---------------------------------------"
            Write-host "Collection Number", $i, "of", $oApp.Sites.Count
     
            if($i -gt 0)
            {
                $mySubweb = $Sites.RootWeb
                $TempRelativeURL = $mySubweb.Url
                ListOrphanedUsers $TempRelativeURL $DomainCNX
            }
        }
     
        Write-Host "======================================="
     
    }
     
    function EnumerateAllSiteColl()
    {
        $farm = Get-SPWebApplication | select DisplayName
        foreach($app in $farm)
        {
            $webapp = Get-SPWebApplication | ? {$_.DisplayName -eq $app.DisplayName}
            Write-Host "Web Application:", $webapp.DisplayName
            ListOrphanedUsersForAllColl $webapp.Url "LDAP://DC=ca,DC=util"
            Write-Host
        }
    }
     
    function StartProcess()
    {
        cls
     
        [System.Diagnostics.Stopwatch] $sw;
        $sw = New-Object System.Diagnostics.StopWatch
        $sw.Start()
     
        EnumerateAllSiteColl
     
        $sw.Stop()
     
        write-host "***************************"
        write-host $GLOBAL:TotalUsersUpdated, "users removed in", $sw.Elapsed.ToString()
        write-host "***************************"
    }
     
    StartProcess


    Enjoy!