Question

class Foo()
{
    public Bar [] bars;
}

class Bar ()
{
    public string name;
    public int id;
}

I have an IEnumerable<Foo> and I want to construct an IEnumerable<Bar> which contains which contains every unique bar (and ordered by ID…but I can handle that part on my own).

I feel that this is the type of task that LINQ is made for..but I can’t seem to figure out how to make it happen.

As an example, if I had:

foo[0] => bars { { name = "one", id = 1 }, { name = "two", id = 2 }, { name = "three", id = 3 },
foo[1] => bars { { name = "four", id = 4 }, { name = "four", id = 4 }

I would want my linq statement to return:

{ name = "one", id = 1 }, { name = "two", id = 2 }, { name = "three", id = 3 }, { name = "four", id = 4 }

Answer 1

If your Bar class overrode Equals and GetHashCode, it would be very simple:

var values = foo.SelectMany(f => f.bars)
                .Distinct()
                .OrderBy(b => b.id);

Without that equality implementation, you’ll need to create your own IEqualityComparer<Foo> to pass into the Distinct method.

If you only need to care about the id to make them distinct, you could use DistinctBy from MoreLINQ to make your life simpler:

var values = foo.SelectMany(f => f.bars)
                .DistinctBy(b => b.id)
                .OrderBy(b => b.id);

Or you could use an anonymous type as a quick hack to get an equality comparer across both values, still with MoreLINQ:

var values = foo.SelectMany(f => f.bars)
                .DistinctBy(b => new { b.id, b.name })
                .OrderBy(b => b.id);

It would almost certainly be better to implement equality in Bar though 🙂