MVVM load data during or after ViewModel construct

问题:

My generic question is as the title states, is it best to load data during ViewModel construction or afterward through some Loaded event handling?

I'm guessing the answer is after construction via some Loaded event handling, but I'm wondering how that is most cleanly coordinated between ViewModel and View?

Here's more details about my situation and the particular problem I'm trying to solve:

I am using the MVVM Light framework as well as Unity for DI. I have some nested Views, each bound to a corresponding ViewModel. The ViewModels are bound to each View's root control DataContext via the ViewModelLocator idea that Laurent Bugnion has put into MVVM Light. This allows for finding ViewModels via a static resource and for controlling the lifetime of ViewModels via a Dependency Injection framework, in this case Unity. It also allows for Expression Blend to see everything in regard to ViewModels and how to bind them.

So anyway, I've got a parent View that has a ComboBox databound to an ObservableCollection in its ViewModel. The ComboBox's SelectedItem is also bound (two-way) to a property on the ViewModel. When the selection of the ComboBox changes, this is to trigger updates in other views and subviews. Currently I am accomplishing this via the Messaging system that is found in MVVM Light. This is all working great and as expected when you choose different items in the ComboBox.

However, the ViewModel is getting its data during construction time via a series of initializing method calls. This seems to only be a problem if I want to control what the initial SelectedItem of the ComboBox is. Using MVVM Light's messaging system, I currently have it set up where the setter of the ViewModel's SelectedItem property is the one broadcasting the update and the other interested ViewModels register for the message in their constructors. It appears I am currently trying to set the SelectedItem via the ViewModel at construction time, which hasn't allowed sub-ViewModels to be constructed and register yet.

What would be the cleanest way to coordinate the data load and initial setting of SelectedItem within the ViewModel? I really want to stick with putting as little in the View's code-behind as is reasonable. I think I just need a way for the ViewModel to know when stuff has Loaded and that it can then continue to load the data and finalize the setup phase.

Thanks in advance for your responses.

回答1:

For events you should use the EventToCommand in MVVM Light Toolkit. Using this you can bind any event of any ui element to relaycommand. Check out his article on EventToCommand at

http://blog.galasoft.ch/archive/2009/11/05/mvvm-light-toolkit-v3-alpha-2-eventtocommand-behavior.aspx

Download the sample and have a look. Its great. You won't need any codebehind then. An example is as follows:

<Page x:Class="cubic.cats.Wpf.Views.SplashScreenView"
      xmlns="http://schemas.microsoft.com/winfx/2006/xaml/presentation"
      xmlns:x="http://schemas.microsoft.com/winfx/2006/xaml"
      xmlns:mc="http://schemas.openxmlformats.org/markup-compatibility/2006" 
      xmlns:d="http://schemas.microsoft.com/expression/blend/2008" 
      xmlns:i="clr-namespace:System.Windows.Interactivity;assembly=System.Windows.Interactivity"
      xmlns:cmd="clr-namespace:GalaSoft.MvvmLight.Command;assembly=GalaSoft.MvvmLight.Extras"
      mc:Ignorable="d" 
      d:DesignHeight="300" d:DesignWidth="300"
    Title="SplashScreenPage">

    <i:Interaction.Triggers>
        <i:EventTrigger EventName="Loaded">
            <cmd:EventToCommand Command="{Binding LoadedCommand}" />
        </i:EventTrigger>        
    </i:Interaction.Triggers>

    <Grid>
        <Label Content="This is test page" />
    </Grid>
</Page>

and the view mode could be like this

 public class SplashScreenViewModel : ViewModelBase
    {
        public RelayCommand LoadedCommand
        {
            get;
            private set;
        }

        /// <summary>
        /// Initializes a new instance of the SplashScreenViewModel class.
        /// </summary>
        public SplashScreenViewModel()
        {
            LoadedCommand = new RelayCommand(() =>
            {
                string a = "put a break point here to see that it gets called after the view as been loaded";
            });
        }
    }

if you would like the view model to have the EventArgs, you can simple set PassEventArgsToCommand to true:

<i:Interaction.Triggers>
            <i:EventTrigger EventName="Loaded">
                <cmd:EventToCommand PassEventArgsToCommand="True" Command="{Binding LoadedCommand}" />
  </i:EventTrigger>        
</i:Interaction.Triggers>

and the view model will be like

public class SplashScreenViewModel : ViewModelBase
{
    public RelayCommand<MouseEventArgs> LoadedCommand
    {
        get;
        private set;
    }

    /// <summary>
    /// Initializes a new instance of the SplashScreenViewModel class.
    /// </summary>
    public SplashScreenViewModel()
    {
        LoadedCommand = new RelayCommand<MouseEventArgs>(e =>
        {
            var a = e.WhateverParameters....;
        });
    }

}

回答2:

Ok, then. :-)

You can bind to a method in the ViewModel by using a behavior.

Here is a link that will help you with that. http://expressionblend.codeplex.com/

回答3:

The following solution is similar to the one already provided and accepted, but it does not use a command in the view model to load the data, but a "normal method". I think commands are more suited for user actions (commands can be available and not available at runtime), that is why a use a regular method call, but also by setting a an interaction trigger in the view.

I suggest this: Create a view model class. Instantiate the view model class within the xaml of the view by creating it inside the DataContext property.

Implement a method to load the data in your view model, e.g. LoadData. Set up the view, so that this method is called when the view loads. This is done by an interaction trigger in your view which is linked to the method in the view model (references to "Microsoft.Expression.Interactions" and "System.Windows.Interactivity" are needed):

View (xaml):

<Window x:Class="MyWindow"
    xmlns="http://schemas.microsoft.com/winfx/2006/xaml/presentation"
    xmlns:x="http://schemas.microsoft.com/winfx/2006/xaml"
    Title="Test" 
    xmlns:viewModel="clr-namespace:ViewModels"
    xmlns:i="http://schemas.microsoft.com/expression/2010/interactivity"
    xmlns:ei="http://schemas.microsoft.com/expression/2010/interactions"            
    >
<Window.DataContext>
    <viewModel:ExampleViewModel/>
</Window.DataContext>
<i:Interaction.Triggers>
    <i:EventTrigger EventName="Loaded">
        <ei:CallMethodAction TargetObject="{Binding}" MethodName="LoadData"/>
    </i:EventTrigger>
</i:Interaction.Triggers>   

This will call the LoadData method in the ViewModel at runtime when the view is loaded. This is where you load your data.

public class ExampleViewModel
{
    /// <summary>
    /// Constructor.
    /// </summary>
    public ExampleViewModel()
    {
        // Do NOT do complex stuff here
    }


    public void LoadData()
    {
        // Make a call to the repository class here
        // to set properties of your view model
    }

If the method in the repository is an async method, you can make the LoadData method async too, but this is not needed in each case.

By the way, generally I would not load data in the constructor of the view model. In the example above the (parameter less) constructor of the view model is called when the designer shows your view. Doing complex things here can cause errors in the designer when showing your view (for the same reason I would not make complex things in the views constructor).

In some scenarios code in the view models constructor can even cause issues at runtime, when the view models constructors executes, set properties of the view model which are bound to elements in the view, while the view object is not completely finished creating.

回答4:

I decided to just have the XAML declaratively bound to a Loaded event handler on the View's code-behind, which in turn just called a method on the ViewModel object, via the View's root element UserControl DataContext.

It was a fairly simple, straight forward, and clean solution. I guess I was hoping for a way to bind the Loaded event to the ViewModel object in the same declarative way you can with ICommands in the XAML.

I may have given Klinger the official answer credit, but he posted a comment to my question, and not an answer. So I at least gave him a one-up on his comment.

回答5:

I had this same problem when dealing with messages between a parent window and a child window. Just change the order in which your view models are created in your ViewModelLocator class. Make sure that all view models which are dependent on a message are created before the view model that sends the message.

For example, in your ViewModelLocator class's constructor:

public ViewModelLocator()
{
    if (s_messageReceiverVm == null)
    {
        s_messageReceiverVm = new MessageReceiverVM();
    }

    if (s_messageSenderVm == null)
    {
        s_messageSenderVm = new MessageSenderVM();
    }
}