Linq Join list with itself distinctly - linq

Question

I have a list of values, say A, B and C. I want to join it with itself to get all distinct combinations considering a value can't join itself and that A-B is the same as B-A
The end result should be
A-B, A-C, B-C
and not a full cross product
A-A, A-B, A-C, B-A, B-B, B-C, C-A, C-B, C-C
In traditional code, I came up with the following
var myList = new List<string> { "A", "B", "C" };
for(var i=0; i < myList.Count; i++)
for (var j=i+1; j < myList.Count; j++)
//Do Something
Is there any LINQ equivalent? I'm mainly seeking a more meaningful and human readable implementation.

Answer 1

I found one way to do it in LInQ but I don't think it qualifies as more meaningful nor readable and mostly, from a performance perspective, it is 3 to 7 (from the first test run I did with only A, B, C in a fixed List) times slower.
from item1 in myList
from item2 in myList.SkipWhile(x => x != item1).Skip(1)
select (item1, item2)
I first thought the main reason the performance was that bad with the LInQ versions was becaus it still ran a full cross product but my simplistic benchmark seems to disagree.
Count
V1
V2
Ratio
V1/item
V2/item
V1/item^2
V2/item^2
1
10
125
455
3,640
12,500
45,500
1,250
4,550
2
10
124
440
3,548
12,400
44,000
1,240
4,400
3
30
976
3034
3,109
32,533
101,133
1,084
3,371
4
30
975
3023
3,101
32,500
100,767
1,083
3,359
5
50
2702
8088
2,993
54,040
161,760
1,081
3,235
6
50
2763
8155
2,952
55,260
163,100
1,105
3,262
7
70
5393
15708
2,913
77,043
224,400
1,101
3,206
8
70
5416
15740
2,906
77,371
224,857
1,105
3,212
9
90
8993
25908
2,881
99,922
287,867
1,110
3,199
10
90
8983
25869
2,880
99,811
287,433
1,109
3,194
Benchmark code:
using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Threading.Tasks;
namespace Benchmarks
{
class Program
{
static void Main(string[] args)
{
for (var i = 1; i <= 10; i+=2)
{
CrossProduct(i, Enumerable.Range(1,10*i).Select(x => x.ToString()).ToList());
}
_ = Console.Read();
}
private static void CrossProduct(int iterations, IList<string> data)
{
var v1 = CrossProductV1(data);
var v2 = CrossProductV2(data);
Console.WriteLine($"{iterations} | DataCount: {data.Count}\t| V1:{v1}* | V2:{v2}");
v2 = CrossProductV2(data);
v1 = CrossProductV1(data);
Console.WriteLine($"{iterations+1} | DataCount: {data.Count}\t| V1:{v1} | V2:{v2}*");
}
private static long CrossProductV1(IList<string> data)
{
var sw = System.Diagnostics.Stopwatch.StartNew();
for (var repeatIndex = 0; repeatIndex < 100000; repeatIndex++)
{
List<object> result = new List<object>();
for (var i = 0; i < data.Count; i++)
for (var j = i + 1; j < data.Count; j++)
result.Add((data[i], data[j]));
}
return sw.ElapsedMilliseconds;
}
private static long CrossProductV2(IList<string> data)
{
var sw = System.Diagnostics.Stopwatch.StartNew();
for (var repeatIndex = 0; repeatIndex < 100000; repeatIndex++)
{
List<object> result = null;
result = (from item1 in data
from item2 in data.SkipWhile(x => x != item1).Skip(1)
select (object)(item1, item2)).ToList();
}
return sw.ElapsedMilliseconds;
}
}
}