LINQ and removing duplicates from an array of objects - linq

Question

I'm trying to de-dupe an array of objects using two columns where the second column is a Dictionary. Best way to describe this is to show some code:
class MyClass
{
public int ID;
public Dictionary<int, int> Dict = new Dictionary<int, int>();
}
And now to create some objects:
List<MyClass> list = new List<MyClass>();
MyClass mc1 = new MyClass();
list.Add(mc1); mc1.ID = 1; mc1.Dict.Add(1, 1);
MyClass mc2 = new MyClass();
list.Add(mc2); mc2.ID = 1; mc2.Dict.Add(1, 1);
MyClass mc3 = new MyClass();
list.Add(mc3); mc3.ID = 1; mc3.Dict.Add(1, 2);
MyClass mc4 = new MyClass();
list.Add(mc4); mc4.ID = 2; mc4.Dict.Add(1, 1);
What I'm looking to accomplish is to distinct by ID and Dict. The results should look like this:
List of MyClass objects (not pretty)
1 //MyClass.ID
1,1 //MyClass.Dictionary
1
1,2
2
1,1
Notice that one of the objects was dropped from the original list because it had a duplicate ID and Dict (dictionary values). I've been playing around with alternate versions of:
var s = from p in list
group p by p.ID into group1
from group2 in
(from p in group1 group p by p.Dict)
group group2 by group1.Key;
but just haven't had any luck. Appreciate any insight folks might have on solving this problem.
PS - I'm not changing the rules but I believe a GROUP BY and a SELECTFIRST will be cleaner than a DISTINCT with its extra code for a Comparer. A pat on the back for anyone who can figure this out using GROUP BY.

Answer 1

For reference types you should add equality comparer in order to do what you want. Add the following class:
public class MyClassComparer : IEqualityComparer<MyClass>
{
public bool Equals(MyClass left, MyClass right)
{
if (left == null && right == null)
{
return true;
}
if (left == null || right == null)
{
return false;
}
if (left.ID == right.ID)
{
if (left.Dict == null && right.Dict == null)
{
return true;
}
if (left.Dict == null || right.Dict == null)
{
return false;
}
if (left.Dict.Count != right.Dict.Count)
{
return false;
}
foreach(var key in left.Dict.Keys)
{
if(!right.Dict.ContainsKey(key))
return false;
if (left.Dict[key] != right.Dict[key])
return false;
}
return true;
}
else return false;
}
public int GetHashCode(MyClass author)
{
return (author.ID).GetHashCode();
}
}
And use that comparer in Distinct override:
List<MyClass> list = new List<MyClass>();
MyClass mc1 = new MyClass();
list.Add(mc1); mc1.ID = 1; mc1.Dict.Add(1, 1);
MyClass mc2 = new MyClass();
list.Add(mc2); mc2.ID = 1; mc2.Dict.Add(1, 1);
MyClass mc3 = new MyClass();
list.Add(mc3); mc3.ID = 1; mc3.Dict.Add(1, 2);
MyClass mc4 = new MyClass();
list.Add(mc4); mc4.ID = 2; mc4.Dict.Add(1, 1);
var result = list.Distinct(new MyClassComparer()).ToList();
You should improve GetHashCode method. It will be your homework :)

Answer 2

Can I have half a pat for the following ? :)
var filteredList = list.GroupBy(mc => mc.ID)
.SelectMany(gr => gr.Distinct(new MyClassComparer()))
.ToList();
Comparer:
public class MyClassComparer : IEqualityComparer<MyClass>
{
public bool Equals(MyClass a, MyClass b)
{
return a.Dict.Count == b.Dict.Count && !a.Dict.Except(b.Dict).Any();
}
public int GetHashCode(MyClass a)
{
return a.ID;
}
}