LINQy way to check if any objects in a collection have the same property value

Question

I have a class Agent with a property Id

Given a collection of Agents I need to check if any of them have duplicate Ids.

I am currently doing this with a hash table but am trying to get Linq-ified, what's a good way of doing this?

Answer 1

foreach(var agent in Agents) {
    if(Agents.Count(a => a.ID == agent.ID) > 1)
        Console.WriteLine("Found: {0}", agent.ID);
}

Answer 2

bool b = list.Any(i => list.Any(j => j.ID == i.ID && j != i));

That's a bit of a brute-force approach but it works. There might be a smarter way to do it using the Except() extension method.

Edit: You didn't actually say that you needed to know which items are "duplicated", only that you needed to know whether any where. This'll do the same thing except give you a list you can iterate over:

list.Where(i => list.Any(j => j.ID == i.ID && j != i))

I like the grouping approach too (group by ID and find the groups with count > 1).

Answer 3

Similar to Y Low's approach,

Edited:

 var duplicates = agents.GroupBy(a => a.ID).Where(a=>a.Count() > 1);

 foreach (var agent in duplicates)
 {
         Console.WriteLine(agent.Key.ToString());
 }

Answer 4

My take (no counting!):

var duplicates = agents
  .GroupBy(a => a.ID)
  .Where(g => g.Skip(1).Any());

Answer 5

For what it's worth, I just compared the two methods we've struck upon in this thread. First I defined a helper class:

public class Foo
{
    public int ID;
}

... and then made a big list of instances with a random ID:

var list = new List<Foo>();

var r = new Random();

for (int i = 0; i < 10000; i++) list.Add(new Foo { ID = r.Next() });

... and lastly, timed the code:

var sw = new Stopwatch();
sw.Start();
bool b = list.Any(i => list.Where(j => i != j).Any(j => j.ID == i.ID));
Console.WriteLine(b);
Console.WriteLine(sw.ElapsedTicks);

sw.Reset();
sw.Start();
b = (list.GroupBy(i => i.ID).Count() != list.Count);
Console.WriteLine(b);
Console.WriteLine(sw.ElapsedTicks);

Here's one output:

False

59392129

False

168151

So I think it's safe to say that grouping and then comparing the count of groups to the count of items is way, way faster than doing a brute-force "nested Any" comparison.