多对多关系基本示例 (MVC3)

Question

I have a MVC3 C# project that I have a model of FoodItem and FoodItemCategory. The two models are shown as follows:

public class FoodItem
{
    public int ID { get; set; }
    [Required]
    public string Name { get; set; }
    public string Description { get; set; }
    public virtual ICollection<FoodItemCategory> Categories { get; set; }
    public DateTime CreateDate { get; set; }
}

public class FoodItemCategory {
    public int ID { get; set; }
    public string Name { get; set; }
    public string Description { get; set; }
    public virtual ICollection<FoodItem> FoodItems { get; set; }
    public DateTime CreateDate { get; set; }
} 

I have a _CreateOrEdit.cshtml view that was initially generated from the Scaffolder and I modified it to include all of the Categories and check the box that the food item belongs to. A food item could have many or all of the categories. The view looks like the following:

@model StackOverFlowIssue.Models.FoodItem
<div class="editor-label">
    @Html.LabelFor(model => model.Name)
</div>
<div class="editor-field">
    @Html.EditorFor(model => model.Name)
    @Html.ValidationMessageFor(model => model.Name)
</div>
<div class="editor-label">
    @Html.LabelFor(model => model.Description)
</div>
<div class="editor-field">
    @Html.EditorFor(model => model.Description)
    @Html.ValidationMessageFor(model => model.Description)
</div>
<div class="editor-label">
    @Html.LabelFor(model => model.Categories)
</div>
<div class="editor-field">
    @foreach (var FoodItemCategory in (IEnumerable<StackOverFlowIssue.Models.FoodItemCategory>)ViewBag.Categories){
        <input type="checkbox" name="FoodItemCategoryId" value="@FoodItemCategory.ID" 
        @foreach(var c in Model.Categories){
            if(c.ID == FoodItemCategory.ID){ 
                @String.Format("checked=\"checked\"")
            } 
        } 
        />
        @FoodItemCategory.Name 
        <br />
    } 
</div>
@Html.Hidden("CreateDate", @DateTime.Now)

As you can see, I have a nested loop that creates a checkbox for each category, and while it is creating each category, I loop through and check for that particular category in the Categories property of my model. If it exists, I set the checked property of the checkbox. If you check a box and click save, on the HttpPost action on the controller, I am performing the following:

    [HttpPost]
    public ActionResult Edit(FoodItem foodItem)
    {
        if (ModelState.IsValid)
        {
            var cList = Request["CategoryId"].Split(',');
            List<FoodItemCategory> categories = new List<FoodItemCategory>();

            foreach (var c in cList) {
                var ci = Convert.ToInt32(c);
                FoodItemCategory category = context.FoodItemCategories.Single(x => x.ID == ci);
                categories.Add(category);
            }

            context.Entry(foodItem).State = EntityState.Modified;
            restaurant.Categories = categories;
            context.SaveChanges();
            return RedirectToAction("Index");
        }
        return View(foodItem);
    }

I am able to save the categories one time. If I go back into the view, and just click save, I receive the following error:

A duplicate value cannot be inserted into a unique index. [ Table name = >FoodItemCategoryFoodItems,Constraint name = PK_ FoodItemCategoryFoodItems _00000000000000A8 ] Description: An unhandled exception occurred during the execution of the current web request. Please >review the stack trace for more information about the error and where it originated in the code.

Exception Details: System.Data.SqlServerCe.SqlCeException: A duplicate value cannot be inserted into >a unique index. [ Table name = FoodItemCategoryFoodItems,Constraint name = >PK_ FoodItemCategoryFoodItems _00000000000000A8 ]

Source Error:

Line 97: context.Entry(foodItem).State = EntityState.Modified; Line 98: foodItem.Categories = categories; Line 99: context.SaveChanges(); Line 100: return RedirectToAction("Index"); Line 101: }

Not sure if it matters but I am using SQLServer Compact Edition 4. Am I going about this the right way? What is the normal coding practice for something like this? I know this same situation occurs daily since this same relationship model is used in many situations like blogs, etc.

Answer 1

Try something like this (untested):

[HttpPost]
public ActionResult Edit(FoodItem foodItem)
{
    if (ModelState.IsValid)
    {
        int id = foodItem.Id;
        // Load food item with related categories first
        var item = context.FoodItems
                          .Include(f => f.Categories)
                          .Single(f => f.Id == id);

        // Process changed scalar values
        context.Entry(item).CurrentValues.SetValues(foodItem);

        // Brute force processing of relations
        // This can be optimized - instead of deleting all and adding all again
        // you can manually compare which relations already exists, add new and
        // remove non existing but let's make that as a homework
        item.Categories.Clear();

        var cList = Request["CategoryId"].Split(',');

        foreach (var c in cList) 
        {
            var ci = Convert.ToInt32(c);
            // Use find - if category was already loaded in the first query, it will
            // be reused without additional query to DB
            var category = context.Categories.Find(ci);
            // Now add category to attached food item to create new relation
            item.Categories.Add(category);
        }

        context.SaveChanges();
        return RedirectToAction("Index");
    }

    return View(foodItem);
}

This can look pretty inefficient but because you are dealing with many-to-many relation where relations can be added or removed in the view it is the only way to do that. The reasons are:

• You must say EF which relations are added and which are deleted
• If you simply add all related categories you are inserting relations again
• You can't say EF which relations are deleted because you don't transfer information about unchecked categories

More about detached object graphs and processing relations is described here . It is about ObjectContext API but DbContext API is just wrapper around that so the same limitations still exist.

Answer 2

In addition to Ladislav's answer, you can get rid of the Request[].split() part by using http://haacked.com/archive/2008/10/23/model-binding-to-a-list.aspx which makes:

[HttpPost]
public ActionResult Edit(FoodItem foodItem, ICollection<int> CategoryId)
{
    if (ModelState.IsValid)
    {
        int id = foodItem.Id;
        // Load food item with related categories first
        var item = context.FoodItems
                          .Include(f => f.Categories)
                          .Single(f => f.Id == id);

        // Process changed scalar values
        context.Entry(item).CurrentValues.SetValues(foodItem);

        // Brute force processing of relations
        // This can be optimized - instead of deleting all and adding all again
        // you can manually compare which relations already exists, add new and
        // remove non existing but let's make that as a homework
        item.Categories.Clear();

        foreach (var id in CategoryID) 
        {
            // Use find - if category was already loaded in the first query, it will
            // be reused without additional query to DB
            var category = context.Categories.Find(id);
            // Now add category to attached food item to create new relation
            item.Categories.Add(category);
        }

        context.SaveChanges();
        return RedirectToAction("Index");
    }

    return View(foodItem);
}