How do I GroupBy from two Lists?

Question

I have two IEnumerable, one with a "Must-have" base of keys, the other with a lot of keys with all kinds of spellings (the same values but containing escape-sequences, different capitalization, etc). I want to create a mapping between every key of the first List to all the ones that correspond with the strings in the second List (Predicate is bool IsPossibleMatch(string a, string b))

So basically I want an IGrouping<string, IEnumerable<string>> because that's exactly the structure of thing I'm looking for.

I tried various versions of keys.GroupBy((x,y) => IsPossibleMatch(x, y)) but couldn't get any of them to work.

Now I'm wondering : Is this even possible with group by? Or do I need to manually do this in some other kind of way?

Please show some example data and also what have you tried? What does it mean "all kinds of spellings"? It is only upper/lower cases differences? — Gilad Green, Apr 08 '19 at 06:02
Upper/Lower, yes, but there are a few other things as well, for example I also have to look out for certain escaped sequences . First List could for example contain "AlphaBeta" and the Grouping would then contain values such as {"AlphaBeta", "alphaBeta", alpha<beta"}, etc. I have already written the predicate for that though: ```bool IsPossibleMatch(string a, string b)``` — Vaethin, Apr 08 '19 at 06:18
Please edit question to include this information in the question — Gilad Green, Apr 08 '19 at 06:18

score 4 · Accepted Answer · answered Apr 08 '19 at 06:22

4

Here is a sample code:

var left = new string[] {
    "key1",
    "key2",
    "key3",
    "key4",
    "key5"
};
var right = new string[]
{
    "key1",
    "Key1",
    "KEy1",
    "KEY1",
    "KeY1",
    "kEY1",
    "kEy1",
    "key2",
    "Key2",
    "KEy2",
    "KEY2",
    "KeY2",
    "kEY2",
    "kEy2"
};
var output = left.GroupJoin(
    right,
    leftStr => leftStr.ToLower(),
    rightStr => rightStr.ToLower(),
    (x,y)=>(x,y)
);
foreach (var leftKey in output)
    Console.WriteLine(leftKey.x + "->" + string.Join(",", leftKey.y));

Produces the following output:

key1->key1,Key1,KEy1,KEY1,KeY1,kEY1,kEy1
key2->key2,Key2,KEy2,KEY2,KeY2,kEY2,kEy2
key3->
key4->
key5->

answered Apr 08 '19 at 06:22

Leighton Ritchie

501
4
15

Group Join! Didn't even know that existed! I figured I had to join but I didn't want to mix up the lists by first joining then Grouping. Thank you! – Vaethin Apr 08 '19 at 06:51
Follow Up Question though: I am trying to modify my existing IsPossibleMatch to implement IEqualityComparer so I can get the right result. It's easy for the Equals(a , b) method but I wonder if I require it at all. Am I guessing correctly that GroupJoin FIRST Calls GetHashCode() to check whether Equals is possible at all and THEN eliminates false-positives (collisions) by calling Equals()? Because for my use I could just make GetHashCode() return the same int every time so I don't have to do the same compute-intensive calculations twice, both in Equals and GetHashCode.@Leighton Richtie – Vaethin Apr 08 '19 at 10:12
Looking at the [reference source](https://referencesource.microsoft.com/#System.Core/System/Linq/Enumerable.cs,c5fc2efafc275a76) shows that it uses the default equality comparer `EqualityComparer.Default`. Perhaps you can try creating your own class with its `Equals` method overriden, then you can make it do whatever inside. However, this is quite micro-optimising and I don't exepct you to gain significant performance improvement. It's more likely that the bottleneck is elsewhere than here. – Leighton Ritchie Apr 09 '19 at 03:19

How do I GroupBy from two Lists?

1 Answers1

Linked