Let's say you have two children; it's trivial to generalize this to any number of children. Fortunately you've already got viewmodels and views, so we're most of the way there. It's just a matter of wiring it all together in a way that works well with WPF.
Here's a set of skeletal viewmodels. Main, and two children. MainViewModel
creates its two child instances. ChildOneViewModel
has a Next button command, bound to a Button
in ChildOneView.xaml.
When the user clicks that button, we want to switch the active view to child two. Rather than have dependencies going in all directions, ChildOneViewModel
is ignorant of what "next" really means; its parent, MainViewModel
, is in charge of that. Everybody knows about his children; you've found that in programming, making a class know about its siblings or its parent creates problems.
So all ChildOneViewModel
does is expose an event so MainViewModel
knows when the button is clicked, and can take any action it likes when that happens. This is cool because what if we could be going to one of two different "next" pages, depending on what the user did in ChildOne? If we move that responsibility to the parent, everything becomes simpler. Easier to write, easier to reuse in a different context.
#region MainViewModel Class
public class MainViewModel : ViewModelBase
{
public MainViewModel()
{
ChildOne = new ChildOneViewModel();
ChildTwo = new ChildTwoViewModel();
ActiveChild = ChildOne;
ChildOne.NextButtonClicked += (s, e) => ActiveChild = ChildTwo;
}
#region ActiveChild Property
private INotifyPropertyChanged _activeChild = default(INotifyPropertyChanged);
public INotifyPropertyChanged ActiveChild
{
get { return _activeChild; }
set
{
if (value != _activeChild)
{
_activeChild = value;
OnPropertyChanged();
}
}
}
#endregion ActiveChild Property
public ChildOneViewModel ChildOne { get; private set; }
public ChildTwoViewModel ChildTwo { get; private set; }
}
#endregion MainViewModel Class
#region ChildOneViewModel Class
public class ChildOneViewModel : ViewModelBase
{
public event EventHandler NextButtonClicked;
// You already know how to implement a command, so I'll skip that.
protected void NextButtonCommandMethod()
{
NextButtonClicked?.Invoke(this, EventArgs.Empty);
}
}
#endregion ChildOneViewModel Class
#region ChildTwoViewModel Class
public class ChildTwoViewModel : ViewModelBase
{
}
#endregion ChildTwoViewModel Class
And here's the XAML that translates all of that into actual UI at runtime:
<Window.Resources>
<!--
These are "implicit datatemplates": They have no x:Key, so they'll be
automatically used to display any content of the specified types.
-->
<DataTemplate DataType="{x:Type local:ChildOneViewModel}">
<local:ChildOneView />
</DataTemplate>
<DataTemplate DataType="{x:Type local:ChildTwoViewModel}">
<local:ChildTwoView />
</DataTemplate>
</Window.Resources>
<Grid>
<!-- other stuff -->
<UserControl
Content="{Binding ActiveChild}"
/>
<!-- other stuff -->
You don't need OpenUserControl()
, FindGrid()
, any of that stuff.