Hello and welcome to the forum.

Enum.group_by is your friend here. Given this list of events you should then do this:

events = [
  %{name: "Item 1", date: ~D[2000-01-01]},
  %{name: "Item 2", date: ~D[2000-01-01]},
  %{name: "Item 3", date: ~D[2000-01-02]},
  %{name: "Item 4", date: ~D[2000-01-03]}
]

Enum.group_by(events, &(&1.date), &(&1.name))

which yields:

%{
  ~D[2000-01-01] => ["Item 1", "Item 2"],
  ~D[2000-01-02] => ["Item 3"],
  ~D[2000-01-03] => ["Item 4"]
}

The Enum.group_by call basically says “group items by the date field and only give me their name field after”.

You can flatten the resulting map to a list exactly the way you need it in the UI. I say give it a try and if you still can’t, I (or somebody else) can give you the next hint.

You could use Enum.chunk_by:

@streams.events
|> Map.values()
|> Enum.sort_by(& &1.date)
|> Enum.chunk_by(& &1.date)
|> Enum.map(fn chunk ->
  # here, chunk is a list of events that all have the same `date` value
  # you can get that date from hd(chunk).date
end)

Can you hide all that Enum mapping/chunking/grouping in a function, passing in the stream, without breaking change tracking? Or does it have to happen in the template to keep change tracking happy?

@al2o3cr @dimitarvp Beautiful, thanks!