LINQ - Full Outer Join

Question

I have a list of people\'s ID and their first name, and a list of people\'s ID and their surname. Some people don\'t have a first name and some don\'t have a surname; I\'d l

Answer 1

Yet another full outer join

As was not that happy with the simplicity and the readability of the other propositions, I ended up with this :

It does not have the pretension to be fast ( about 800 ms to join 1000 * 1000 on a 2020m CPU : 2.4ghz / 2cores). To me, it is just a compact and casual full outer join.

It works the same as a SQL FULL OUTER JOIN (duplicates conservation)

Cheers ;-)

using System;
using System.Collections.Generic;
using System.Linq;
namespace NS
{
public static class DataReunion
{
    public static List> FullJoin(List List1, Func KeyFunc1, List List2, Func KeyFunc2)
    {
        List> result = new List>();

        Tuple[] identifiedList1 = List1.Select(_ => Tuple.Create(KeyFunc1(_), _)).OrderBy(_ => _.Item1).ToArray();
        Tuple[] identifiedList2 = List2.Select(_ => Tuple.Create(KeyFunc2(_), _)).OrderBy(_ => _.Item1).ToArray();

        identifiedList1.Where(_ => !identifiedList2.Select(__ => __.Item1).Contains(_.Item1)).ToList().ForEach(_ => {
            result.Add(Tuple.Create(_.Item2, default(T2)));
        });

        result.AddRange(
            identifiedList1.Join(identifiedList2, left => left.Item1, right => right.Item1, (left, right) => Tuple.Create(left.Item2, right.Item2)).ToList()
        );

        identifiedList2.Where(_ => !identifiedList1.Select(__ => __.Item1).Contains(_.Item1)).ToList().ForEach(_ => {
            result.Add(Tuple.Create(default(T1), _.Item2));
        });

        return result;
    }
}
}

The idea is to

1. Build Ids based on provided key function builders
2. Process left only items
3. Process inner join
4. Process right only items

Here is a succinct test that goes with it :

Place a break point at the end to manually verify that it behaves as expected

using System;
using System.Collections.Generic;
using Microsoft.VisualStudio.TestTools.UnitTesting;
using Newtonsoft.Json;
using Newtonsoft.Json.Linq;
using NS;

namespace Tests
{
[TestClass]
public class DataReunionTest
{
    [TestMethod]
    public void Test()
    {
        List> A = new List>();
        List> B = new List>();

        Random rnd = new Random();

        /* Comment the testing block you do not want to run
        /* Solution to test a wide range of keys*/

        for (int i = 0; i < 500; i += 1)
        {
            A.Add(Tuple.Create(rnd.Next(1, 101), rnd.Next(1, 101), "A"));
            B.Add(Tuple.Create(rnd.Next(1, 101), rnd.Next(1, 101), "B"));
        }

        /* Solution for essential testing*/

        A.Add(Tuple.Create(1, 2, "B11"));
        A.Add(Tuple.Create(1, 2, "B12"));
        A.Add(Tuple.Create(1, 3, "C11"));
        A.Add(Tuple.Create(1, 3, "C12"));
        A.Add(Tuple.Create(1, 3, "C13"));
        A.Add(Tuple.Create(1, 4, "D1"));

        B.Add(Tuple.Create(1, 1, "A21"));
        B.Add(Tuple.Create(1, 1, "A22"));
        B.Add(Tuple.Create(1, 1, "A23"));
        B.Add(Tuple.Create(1, 2, "B21"));
        B.Add(Tuple.Create(1, 2, "B22"));
        B.Add(Tuple.Create(1, 2, "B23"));
        B.Add(Tuple.Create(1, 3, "C2"));
        B.Add(Tuple.Create(1, 5, "E2"));

        Func, Tuple> key = (_) => Tuple.Create(_.Item1, _.Item2);

        var watch = System.Diagnostics.Stopwatch.StartNew();
        var res = DataReunion.FullJoin(A, key, B, key);
        watch.Stop();
        var elapsedMs = watch.ElapsedMilliseconds;
        String aser = JToken.FromObject(res).ToString(Formatting.Indented);
        Console.Write(elapsedMs);
    }
}

}