Question

I have the following array:

Driver[] predictions = new Driver[6];
predictions[0] = new Driver(10, "Michael Schumacher");
predictions[1] = new Driver(10, "Michael Schumacher");
predictions[2] = new Driver(9, "Fernando Alonso");
predictions[3] = new Driver(8, "Jensen Button");
predictions[4] = new Driver(7, "Felipe Massa");
predictions[5] = new Driver(6, "Giancarlo Fisichella");

I want to get all the duplicates – The name once, and then the positions (index) where the duplicates are. So, in this case I want to get “Michael Schumacher” and the positions 1 and 2 (index 0 and 1).

Can this be done in one go, or do I need to consider other options? I just read on DotNetPearls that IndexOf is pretty slow compared to having your own logic.

var driversSelectedMoreThanOnceAndTheirPositions = predictions.Select((driver, index) => new { driver, index })
.GroupBy(item => item.driver.Name)
.Where(grp => grp.Count() > 1)
.ToDictionary(g => g.Key, g => g.Select(a => (a.index + 1)).ToList());

Answer 1

To utilize linq for this, you could write something like the following, which uses an overload of Select that allows you to get the element index and then performs a GroupBy operation.

var query = 
 predictions.Select((driver, index) => new { driver, index })
 .GroupBy(item => item.driver.Name)
 .Where(grp => grp.Count() > 1)
 .Select(grp => new { Name = grp.Key, Indexes = grp.Select(item => item.index) });

This will result in a sequence of anonymous-typed objects with the properties

class Anon
{
    public string Name;
    public IEnumerable<int> Indexes;
}

Which you could use as in

foreach (var item in query)
{
    Console.WriteLine(item.Name);
    foreach (int index in item.Indexes)
        Console.WriteLine(index);
}

You can, of course, change the way you group to get the entire driver object.