Question

I know this is a little bit hypothetical, because I am not looping this, it only occurs once or twice in a program run, and so it will only be a completely unnoticable amount of time, but I want to know if one of these is better than the other, or if it doesn’t matter at all, because the optimiser will optimise away the worse code.

I have this class:

class FOO_Data
{
    private string description, url;
    private bool editable;

    public FOO_Data(string description, string url, bool editable)
    {
        this.description = description;
        this.url = url;
        this.editable = editable;
    }

    public string Description { get { return description; } set { description = value; } }
    public string URL { get { return url; } set { url = value; } }
    public bool Editable { get { return editable; } set { editable = value; } }
} 

At some other point in my code, I need to edit an instance of this class in an array of this class.

Which one is better? This one:

array[index] = new FOO_Data(data.Description, data.URL, System.Convert.ToBoolean(data.Editable));

or this one:

array[index].Description = data.Description;
array[index].Editable = Convert.ToBoolean(data.Editable);
array[index].URL = data.URL;

I would tend towards the first, but I am not very sure. I would appreciate any insight you could offer.

Thanks a lot!

Answer 1

If array[index] is null, the second bit of code, trying to access members will throw a NullReferenceException.

This can’t happen with the first bit of code, where you are assigning a newly constructed FOO_Data object.

In terms of performance, you are unlikely to see any difference if the array is indeed fully populated as object creation is a very light process.