Question

I am new to generics in C# and while reading a book stumbled upon an example:

var cars = from car in data.AsEnumerable()
  where
    car.Field<string>("Color") == "Red"
    select new
    {
      ID = car.Field<int>("CarID"),
      Make = car.Field<string>("Make")
    };

The author says that car.Field<string>("Color") gives the additional compile-time checking comparing to (string)car["Color"]. But how does the compiler know that car.Field<string>("Color") is compilable for “Color” and not for “CarID”? Or there is some kind of another “additional compile-time checking” that I miss?

Answer 1

It doesn’t give you any additional compile-time checking. If you use the wrong type, in both cases you’ll get an exception during run-time.

But it can be useful to do additional stuff that simple cast can’t. For example Field<int>("CarId") could call a method that converts the string in the field to an int.

And assuming you’re talking about DataRow.Field<T>(), then, according to the documentation, it’s useful mostly for dealing with null values and nullable types correctly.