Question

Does anyone know about a good way to accomplish this task?

Currently i’m doing it more ore less this way, but i’m feeling someway unhappy with this code, unable to say what i could immediately improve.

So if anyone has a smarter way of doing this job i would be happy to know.

private bool Check(List<MyItem> list)
{
    bool result = true;
    //MyItem implements IComparable<MyItem>
    list.Sort();

    for (int pos = 0; pos < list.Count - 1; pos++)
    {
        bool previousCheckOk = true;
        if (pos != 0)
        {
            if (!CheckCollisionWithPrevious(pos))
            {
                MarkAsFailed(pos);
                result = false;
                previousCheckOk = false;
            }
            else
            {
                MarkAsGood(pos);
            }
        }

        if (previousCheckOk && pos != list.Count - 1)
        {
            if (!CheckCollisionWithFollowing(pos))
            {
                MarkAsFailed(pos);
                result = false;
            }
            else
            {
                MarkAsGood(pos);
            }
        }
    }
    return result;
}

private bool CheckCollisionWithPrevious(int pos)
{
    bool checkOk = false;
    var previousItem = _Item[pos - 1];

    // Doing some checks ...

    return checkOk;
}

private bool CheckCollisionWithFollowing(int pos)
{
    bool checkOk = false;
    var followingItem = _Item[pos + 1];

    // Doing some checks ...

    return checkOk;
}

Update

After reading the answer from Aaronaught and a little weekend to refill full mind power i came up with the following solution, that looks far better now (and nearly the same i got from Aaronaught):

public bool Check(DataGridView dataGridView)
{
    bool result = true;
    _Items.Sort();

    for (int pos = 1; pos < _Items.Count; pos++)
    {
        var previousItem = _Items[pos - 1];
        var currentItem = _Items[pos];

        if (previousItem.CollidesWith(currentItem))
        {
            dataGridView.Rows[pos].ErrorText = "Offset collides with item named " + previousItem.Label;
            result = false;
            sb.AppendLine("Line " + pos);
        }
    }

    dataGridView.Refresh();
    return result;
}

Answer 1

It’s certainly possible to reduce the repetition:

private bool Check(List<MyItem> list)
{
    list.Sort();

    for (int pos = 1; pos < list.Count; pos++)
    {
        if (!CheckCollisionWithPrevious(list, pos))
        {
            MarkAsFailed();
            return false;
        }
        MarkAsGood();
    }
    return true;
}

private bool CheckCollisionWithPrevious(List<MyItem> list, int pos)
{
    bool checkOk = false;
    var previousItem = list[pos - 1];

    // Doing some checks ...

    return checkOk;
}

Assuming that CheckCollisionWithPrevious and CheckCollisionWithFollowing perform essentially the same comparisons, then this will perform the same function with a lot less code.

I’ve also added the list as a parameter to the second function; it doesn’t make sense to be taking it as a parameter in the first function, but then referencing a hard-coded member in the function it calls. If you’re going to take a parameter, then pass that parameter down the chain.

As far as performance is concerned, though, you’re re-sorting the list every time this happens; if it happens often enough, you might be better off using a sorted collection to begin with.

Edit: And just for good measure, if the whole point of this code is just to check for some kind of duplicate key, then you would be way better off using a data structure that prevents this in the first place, such as a Dictionary<TKey, TValue>.