Question

I have an in memory List<TreeNode>. Is it possible to to use linq to select all leaf nodes?

The SQL would look something like this:

SELECT      
    t1.Name
FROM 
    Tree t1 
    LEFT JOIN Tree t2 
        ON t1.UserId = t2.ParentId  
WHERE 
    t2.UserId is null

The closest that I’ve got is (i still need to filter out the nulls):

                       var test = from t1 in list
                       join t2 in list
                       on t1.UserId equals t2.ParentId into g                   

                       from result in g.DefaultIfEmpty()          

                       select new { 
                       t1.Name
                       };

Which produces the correct sql (when testing in linqpad) without the where clause. I can’t access t2 where I select the anonymous object of if I try to filter results on where t2 == null. Do I need to create a copy of the list and join onto that?

Answer 1

EDIT: Okay, I’d misunderstood before. So you’re basically after items (t1) in the list where there isn’t no element t2 such that t1.UserId == t2.ParentId? In that case, I suggest you use:

I suspect you want:

var test = from t1 in list
           join t2 in list
              on t1.UserId equals t2.ParentId into g
           where !g.Any()
           select t1.Name;

Note that I’m not using an anonymous type here as you’re only selecting a single value.

Another alternative would be:

var parents = new HashSet<Guid>(list.Where(x => x.ParentId != null)
                                    .Select(x => x.ParentId.Value));
var query = list.Where(t1 => !parents.Contains(t1.UserId));