Question

Given the following method:

public static void SetPropertyValue(object target, string propName, object value)
{
    var propInfo = target.GetType().GetProperty(propName,
                         BindingFlags.Instance | BindingFlags.Public | BindingFlags.NonPublic | BindingFlags.DeclaredOnly);

    if (propInfo == null)
        throw new ArgumentOutOfRangeException("propName", "Property not found on target");
    else
        propInfo.SetValue(target, value, null);
}

How would you go about writing it’s expression enabled equivalent without needing to pass in an extra parameter for target?

Why do this instead of setting the property directly I can hear you say. For example suppose we have the following class with a property that has a public getter but private setter:

public class Customer 
{
   public string Title {get; private set;}
   public string Name {get; set;}
}

I would like to be able to call:

var myCustomerInstance = new Customer();
SetPropertyValue<Customer>(cust => myCustomerInstance.Title, "Mr");

Now here is some sample code.

public static void SetPropertyValue<T>(Expression<Func<T, Object>> memberLamda , object value)
{
    MemberExpression memberSelectorExpression;
    var selectorExpression = memberLamda.Body;
    var castExpression = selectorExpression as UnaryExpression;

    if (castExpression != null)
        memberSelectorExpression = castExpression.Operand as MemberExpression;
    else
        memberSelectorExpression = memberLamda.Body as MemberExpression;

    // How do I get the value of myCustomerInstance so that I can invoke SetValue passing it in as a param? Is it possible

}

Any pointers?

Answer 1

You could cheat and make life easier with an extension method:

public static class LambdaExtensions
{
    public static void SetPropertyValue<T, TValue>(this T target, Expression<Func<T, TValue>> memberLamda, TValue value)
    {
        var memberSelectorExpression = memberLamda.Body as MemberExpression;
        if (memberSelectorExpression != null)
        {
            var property = memberSelectorExpression.Member as PropertyInfo;
            if (property != null)
            {
                property.SetValue(target, value, null);
            }
        }
    }
}

and then:

var myCustomerInstance = new Customer();
myCustomerInstance.SetPropertyValue(c => c.Title, "Mr");

The reason why this is easier is because you already have the target on which the extension method is invoked. Also the lambda expression is a simple member expression without closures. In your original example the target is captured in a closure and it could be a bit tricky to get to the underlying target and PropertyInfo.