Binding to a collection of CheckboxFor

问题:

I'm trying to implement a permission screen in which a user can be given a particular permission on a particular screen. For this I'm generating a collection of Checkboxfor, bound to a collection of bool properties. But when I submit the form, I'm either getting all bool properties true or all false, depending on whether I initialized these properties as true or false in the viewmodel constructor.

Here is the code for the ViewModel:

Approach I:

public class MyViewModel
{
    public MyModel Model { get; set; }        

    public IEnumerable<ScreenType> Screens { get; set; }
    public IEnumerable<SecurityType> SecurityTypes { get; set; }
    public List<PermissionType> Permissions { get; set; }        

    public MyViewModel()
    {
        LoadScreens();
        LoadSecurityTypes();
        LoadPermissions();
    }

    public void LoadPermissions()
    {
        Permissions = new List<PermissionType>();

        foreach (var screen in Screens)
        {
            foreach (var securityType in SecurityTypes)
            {
                Permissions.Add(
                    new PermissionType
                    {
                        PermissionId= Guid.NewGuid(),
                        ScreenId= screen.Id,
                        SecurityId = securityType.Id,
                        IsAllowed = false
                    });
            }
        }
    }    
}

Approach II

public class MyViewModel
{
    public MyModel Model { get; set; }        

    public IEnumerable<ScreenType> Screens { get; set; }
    public IEnumerable<SecurityType> SecurityTypes { get; set; }
    public List<bool> AllowedList { get; set; }        

    public MyViewModel()
    {
        LoadScreens();
        LoadSecurityTypes();
        LoadPermissions();
    }

    public void LoadPermissions()
    {
        AllowedList = new List<bool>();

        foreach (var form in Forms)
        {
            foreach (var security in Securities)
            {
                AllowedList.Add(false);
            }
        }
    }    
}

Here is the code my the view:

Approach I:

    @using (Ajax.BeginForm("Create", "Role", null, new AjaxOptions { UpdateTargetId = "addStatus", InsertionMode = InsertionMode.Replace, OnSuccess = "onFormPostSuccess" }, new { @id = "AddForm" }))
    {  
        <div>
            <span><label>Screen</label></span>
            @foreach (var security in Model.SecurityTypes)
            { 
                <span><label>@security.Name</label></span>                    
            }
            <br />
            @foreach (var screen in Model.Screens)
            {
                <span>@screen.Name</span>
                foreach (var security in Model.SecurityTypes)
                { 
                    <span>@Html.CheckBoxFor(m => m.Permissions.Where(s => s.SecurityId == security.Id && s.ScreenId == screen.Id).Single().IsAllowed, new { @id = HtmlHelper.GenerateIdFromName("Create" + screen.Name + security.Name) })</span>                   
                }
                <br />
            }
        </div>
        <div>
            <span>
                <input type="image" src="/content/images/submit_button.png" value="submit" />
            </span> 
        </div>
        <div>
            <span id="addStatus" class="submitStatus"></span>
        </div>
    }

Approach II:

 @using (Ajax.BeginForm("Create", "Role", null, new AjaxOptions { UpdateTargetId = "addStatus", InsertionMode = InsertionMode.Replace, OnSuccess = "onFormPostSuccess" }, new { @id = "AddForm" }))
    {  
        <div>
            <span><label>Screen</label></span>
            @foreach (var security in Model.SecurityTypes)
            { 
                <span><label>@security.Name</label></span>                    
            }
            <br />
            @foreach (int i=0; i < Model.Screens.Count(); i++)
            {
                <span>@Model.Screens.ElementAt(i).Name</span>
                foreach (int j=0; j<Model.SecurityTypes.Count(); j++)
                { 
                    @* The index 5*i+j is because I have 5 security types *@ 
                    <span>@Html.CheckBoxFor(Model.AllowedList[5*i+j], new { @id = HtmlHelper.GenerateIdFromName("Create" + @Model.Screens.ElementAt(i).Name + @Model.SecurityTypes.ElementAt(j).Name) })</span>                   
                }
                <br />
            }
        </div>
        <div>
            <span>
                <input type="image" src="/content/images/submit_button.png" value="submit" />
            </span> 
        </div>
        <div>
            <span id="addStatus" class="submitStatus"></span>
        </div>
    }

Here's the code for the Create actionmethod in Controller:

    [Authorize]
    [HttpPost]
    public JsonResult Create(MyViewModel viewModel)
    {   
        if ( ModelState.IsValid)
        {                

            if (service.AddRole(viewModel))
            {                    
                return Json("Role Added !");
            }
            return Json("Role exists !");
        }
        return Json("Please correct errors");
    }

When I check the viewModel in the Create actionmethod, all the IsAllowed properties are false. As were initialized in the viewmodel contructor. There is no effect of checking/unchecking Checkboxes from the view.

Am I missing something?

回答1:

Strongly typed HTML helpers such as the CheckBoxFor work only with simple expressions as first argument (property and array index access at most). The following expression is something completely out of the capacities of this helper:

m => m.Permissions.Where(s => s.SecurityId == security.Id && s.ScreenId == screen.Id).Single().IsAllowed

I would recommend you to use a real view model and perform this at your controller level when mapping between your domain model and the view model.

---

UPDATE:

Apparently my answer about view models wasn't clear enough, so let me try to exemplify what I mean by view models.

So we start by defining our view models that will be required for implementing the required logic for this view:

public class CreateViewModel
{
    public IEnumerable<ScreenViewModel> Screens { get; set; }
}

public class ShowCreateViewModel: CreateViewModel
{
    public IEnumerable<SecurityTypeViewModel> SecurityTypes { get; set; }
}

public class ScreenViewModel
{
    public string ScreenName { get; set; }
    [HiddenInput(DisplayValue = false)]
    public int ScreenId { get; set; }
    public IEnumerable<SecurityTypeViewModel> SecurityTypes { get; set; }
}

public class SecurityTypeViewModel
{
    public string SecurityName { get; set; }
    [HiddenInput(DisplayValue = false)]
    public int SecurityTypeId { get; set; }
    public bool IsAllowed { get; set; }
}

Then we could have a controller action that will take care of fetching the domain models from a repository or something and map to the view models:

public class HomeController : Controller
{
    public ActionResult Create()
    {
        // The information will obviously come from a domain model that 
        // we map to a view model, but for the simplicity of the answer
        // I am hardcoding the values here
        var securityTypes = new[]
        {
            new SecurityTypeViewModel { SecurityTypeId = 1, SecurityName = "security 1" },
            new SecurityTypeViewModel { SecurityTypeId = 2, SecurityName = "security 2" },
            new SecurityTypeViewModel { SecurityTypeId = 3, SecurityName = "security 3" },
        };

        // The information will obviously come from a domain model that 
        // we map to a view model, but for the simplicity of the answer
        // I am hardcoding the values here
        return View(new ShowCreateViewModel
        {
            SecurityTypes = securityTypes,
            Screens = new[]
            {
                new ScreenViewModel 
                {
                    ScreenId = 1,
                    ScreenName = "Screen 1",
                    SecurityTypes = securityTypes
                },
                new ScreenViewModel 
                {
                    ScreenId = 2,
                    ScreenName = "Screen 2",
                    SecurityTypes = securityTypes
                },
            }
        });
    }

    [HttpPost]
    public ActionResult Create(CreateViewModel model)
    {
        // The view model passed here will contain all the necessary information
        // for us to be able to perform the actual Save: 
        // a list of the screen ids along with a list of the selected permission ids

        return Content(
            "Thank you for selecting the following allowed permissions:<br/>" + 
            string.Join("<br/>", model.Screens.Select(s => string.Format(
                "screen id: {0}, permission ids: {1}", 
                s.ScreenId, 
                string.Join(",", s.SecurityTypes.Where(st => st.IsAllowed).Select(st => st.SecurityTypeId))
            )))
        );
    }
}

and now what's left is to define the view and the corresponding editor/display templates.

Let's start with the main view (~/Views/Home/Create.cshtml):

@model ShowCreateViewModel

<script src="@Url.Content("~/Scripts/jquery.unobtrusive-ajax.js")" type="text/javascript"></script>

@using (Ajax.BeginForm("Create", "Home", null, new AjaxOptions { UpdateTargetId = "addStatus", InsertionMode = InsertionMode.Replace, OnSuccess = "onFormPostSuccess" }, new { @id = "AddForm" }))
{  
    <table>
        <thead>
            <tr>
                <th>Screen/Security type</th>
                @Html.DisplayFor(x => x.SecurityTypes)
            </tr>
        </thead>
        <tbody>
            @Html.EditorFor(x => x.Screens)
        </tbody>
    </table>    

    <div>
        <input type="submit" value="submit" />
    </div>

    <div id="addStatus" class="submitStatus"></div>
}

Next we have an editor template for the ScreenViewModel model (~/Views/Shared/EditorTemplates/ScreenViewModel.cshtml):

@model ScreenViewModel
<tr>
    <td>
        @Html.DisplayFor(x => x.ScreenName)
        @Html.EditorFor(x => x.ScreenId)
    </td>
    @Html.EditorFor(x => x.SecurityTypes)
</tr>

Then the editor template for the SecurityTypeViewModel model (~/Views/Shared/EditorTemplates/SecurityTypeViewModel.cshtml):

@model SecurityTypeViewModel
<td>
    @Html.CheckBoxFor(x => x.IsAllowed)
    @Html.EditorFor(x => x.SecurityTypeId)
</td>

And finally the display template for the SecurityTypeViewModel model (~/Views/Shared/DisplayTemplates/SecurityTypeViewModel.cshtml):

@model SecurityTypeViewModel
<th>
    @Html.DisplayFor(x => x.SecurityName)
</th>

And that's pretty much it:

I have left for you the mapping between your actual domain models and the view models defined here.