How to handle duplicate keys while joining two lists?

Question

I'm new to C#.

I have the following struct.

struct Foo
{
    string key;
    Bar values;
}

I have two lists of Foo, L1 and L2 of equal size both contain same set of keys.

I have to merge the corresponding Foo instances in L1 and L2.

Foo Merge(Foo f1, Foo f2)
{
   // merge f1 and f2.
   return result.
}

I wrote the following to achieve this.

resultList = L1.Join(L2, f1 => f1.key, f2 => f2.key, (f1, f2) => Merge(f1, f2)
                        ).ToList())

My problem is that my key is not unique. I have n number of elements in L1 with the same key (say "key1") (which are also appearing in L2 somewhere). So, the above join statement selects n matching entries from L2 for each "key1" from L1 and I get n*n elements with key "key1" in the result where I want only n. (So, this is kind of crossproduct for those set of elements).

I want to use Join and still select an element from L1 with "key1" and force the Linq to use the first available 'unused' "key1" element from L2. Is this possible? Is join a bad idea here?

(Also, the I want to preserve the order of the keys as in L1. I tried to handle all elements with such keys before the join and removed those entries from L1 and L2. This disturbed the order of the keys and it looked ugly).

I'm looking for a solution without any explicit for loops.

Answer 1

From your comment to ElectricRouge answer, you could do something like

var z = list1.Join(list2.GroupBy(m => m.Id), 
                   m => m.Id, 
                   g => g.Key, 
                   (l1, l2) => new{l1, l2});

this would give you a list of all keys in l1, and the corresponding grouped keys in l2.

Not sure it's really readable.

Answer 2

I need to find the corresponding entries in two lists and do some operation on them. That is my preliminary requirement.

For this you can do something like this.

var z=S1.Select(i=>i.Key).Tolist(); //make a list of all keys in S1
List<Foo> result=new List<Foo>();
foreach(var item in z)    // Compare with S2 using keys in z
{
   var x=item.Where(i=>i.Key==item.Key)
   result.Add(x);  
}

Is this what you are looking for?

Answer 3

I want to use Join and still select an element from L1 with "key1" and force the Linq to use the first available 'unused' "key1" element from L2. Is this possible?

When combining elements from the two lists you want to pick the first element in the second list having the same key as the element in the first list. (Previously, I interpreted you question differently, and a solution to this different problem is available in the edit history of this answer.)

For quick access to the desired values in the second list a dictionary is created providing lookup from keys to the desired value from the second list:

var dictionary2 = list2
  .GroupBy(foo => foo.Key)
  .ToDictionary(group => group.Key, group => group.First());

The use of First expresses the requirement that you want to pick the first element in the second list having the same key.

The merged list is now created by using projection over the first list:

var mergedList = list1.Select(
  foo => Merge(
    foo,
    dictionary2[foo.Key]
  )
);

When you use foreach to iterate mergedList or ToList() the desired result will be computed.

Answer 4

You could use Union to remove the duplicated keys. Documentation at http://msdn.microsoft.com/en-us/library/bb341731.aspx

List<int> list1 = new List<int> { 1, 12, 12, 5};
List<int> list2 = new List<int> { 12, 5, 7, 9, 1 };
List<int> ulist = list1.Union(list2).ToList();

Example taken from : how to merge 2 List<T> with removing duplicate values in C#

Or you can use Concat to merge a list of different types (Keeping all keys). See the documentation her : http://msdn.microsoft.com/en-us/library/bb302894(v=vs.110).aspx

var MyCombinedList = Collection1.Concat(Collection2)
                                .Concat(Collection3)
                                .ToList();

Example taken from same question : Merge two (or more) lists into one, in C# .NET

Answer 5

Finally I adapted Raphaël's answer as below.

public class FooComparer : IEqualityComparer<Foo>
{
      public bool Equals(Foo o1, Foo o2)
      {
          return o1.key == o2.key;
      }

      public int GetHashCode(Foo obj)
      {
          return obj.key.GetHashCode();
      }
}


resultList = L1.Join(L2.Select(m => m).Distinct(new FooComparer()).ToList(), f1 => f1.key, f2 => f2.key, (f1, f2) => Merge(f1, f2)
                        ).ToList());

Short explanation:

L2.Select(m => m).Distinct(new FooComparer()).ToList()

creates a new list by removing the duplicate keys from L2. Join L1 with this new list to get the required result.