The reason why such operations doesn't exist on stream is because streams
may encapsulate resources. For example, Repo.stream is a lazy database
query. Having operations such as Stream.head(...) | Stream.tail(...) means
you would have to execute the query twice, one to get the head, and another
for the tail, which can be very wasteful in those cases.
You probably could be tackled without head and tail, by first applying all
transforms to streams and then running it:
~e[*/10]
|> Crontab.Scheduler.get_next_run_dates()
|> Stream.each(fn date ->
date
|> DateTime.from_naive!("Etc/UTC")
|> DateTime.diff(DateTime.utc_now(), :millisecond)
|> min(0)
|> Process.sleep()
callback.(date)
end)
|> Stream.run()
The stream programming model requires you to think about working with the
collection as a whole, instead of item by item.
On Thu, Jun 11, 2020 at 4:10 PM Jonatan Männchen <[email protected]>
wrote:
> I would like to propose two new functions for the `Stream` module:
>
> *Head Function*
>
> @spec head(Enumerable.t) :: Stream.element
>
> This function is the equivalent of the following code:
>
> [element] = stream
> |> Stream.take(1)
> |> Enum.to_list
>
> *Tail Function*
>
> @spec tail(Enumerable.t) :: Enumerable.t # %Stream{...}
>
> This function returns a new stream without the head element.
>
> *Syntax Sugar*
>
> It would further be awesome if the following would work:
>
> [element1, elementN... | rest] = stream # I'm aware that this would not
> work as a function head / case etc.
> %Stream{} = rest
>
> *Real Wold Example*
>
> This capability would for example enable the following (using
> https://hexdocs.pm/crontab):
>
> def setup do
> ~e[*/10] # Every 10 Minutes
> |> Crontab.Scheduler.get_next_run_dates() # Creating a Stream with
> dates matching schedule
> |> run(fn date ->
> # Do something useful
> end)
> end
>
>
> def run(schedules, callback) do
> [next_date | rest] = schedules # With Syntax Sugar
>
> next_date = Stream.head(schedules) # Without Syntax Sugar
> rest = Stream.tail(schedules) # Without Syntax Sugar
>
> next_date
> |> DateTime.from_naive!("Etc/UTC")
> |> DateTime.diff(DateTime.utc_now(), :millisecond)
> |> min(0)
> |> Process.sleep()
>
> callback.(next_date)
>
> run(schedules, callback)
> end
>
> *Problems*
>
> There's some problems with this exact approach since pattern matching
> doesn't work with streams. Therefore all of the above will need some work.
> It should however illustrate the intention of this proposal.
>
> --
> You received this message because you are subscribed to the Google Groups
> "elixir-lang-core" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/elixir-lang-core/31cc6e90-2235-4f31-aa76-9fc045b8aa88o%40googlegroups.com
> <https://groups.google.com/d/msgid/elixir-lang-core/31cc6e90-2235-4f31-aa76-9fc045b8aa88o%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
>
--
You received this message because you are subscribed to the Google Groups
"elixir-lang-core" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/elixir-lang-core/CAGnRm4%2Bq%2BYdWxw2npc4wbah%2Bu3h8uXtdQYQd%2BS78OmfPrNVPJw%40mail.gmail.com.
