The first thing we want is to use Flow to count how many files each user has currently stored on ElixirDrip. In practice, we want to reduce our media_set, by incrementing each user's counter for each file. The naive approach could be something like the following:
$ cat examples/search.exsimport_file("./examples/sample_data.exs")# ...files_by_user_v1 = media_set |> Flow.from_enumerable(max_demand: 1) |> Flow.reduce(fn -> %{} end, fn media, accum -> Map.update(accum, media.user_id, 1, &(&1 + 1)) end)
However, if we run the script and check the outcome of the files_by_user_v1 flow, we will see a weird result:
# 'search.exs' already evaluatediex> files_by_user_v1 |> Enum.to_list()[{1, 1}, {3, 1}, {2, 1}, {3, 2}, {4, 1}]
The explanation ...
