I like Bruce's option a lot and I will proceed with that. Thanks Bruce and
everyone!
On Wed, Dec 9, 2020 at 1:48 AM Anil Kulkarni wrote:
> If there was no Enum.with_index/2 function in existence today, how would
> we implement it? My personal preference is to do {index, value}, as per
> José’s
Hi Blake, Logger already provides a custom formatter and you can replace
the formatter to emit JSON instead.
We don't intend to include a JSON specific formatter by default because
this is something that can be trivially done as a library (and there is no
benefit doing it from core itself
Considered the growth of structured logging, it would be really nice if
elixir included a JSON Logger backend by default. This would be very basic
and not include any weird paths or third party service specific fields, but
it would wrap up multi line outputs (like stack traces) to a single line
You could also use the new zip_with:
Map.new(Enum.zip_with(list, 0..length(list), & {&2, &1}))
On Tue, 8 Dec 2020 at 23:51, Christopher Keele
wrote:
> I think that even if it enabled awesome optimization opportunities,
> 1. Having both *Enum.index_with* and *Enum.index*/*indexed* will confuse
I think that even if it enabled awesome optimization opportunities,
1. Having both *Enum.index_with* and *Enum.index*/*indexed* will confuse
people.
2. Straying from Elixir's consistent API design of "Module X functions take
X's as their first argument" will confuse people.
a. Especially in
There is a bit of a collision with the option for the offset, but you could
pattern match that.
-bt
On Tue, Dec 8, 2020 at 4:57 PM Zach Daniel
wrote:
> I like that idea a lot Bruce. Let’s you do something like:
>
> maps |> Enum.with_index((&1, :index, &2))
>
> On Tue, Dec 8, 2020 at 4:50 PM
I like that idea a lot Bruce. Let’s you do something like:
maps |> Enum.with_index((&1, :index, &2))
On Tue, Dec 8, 2020 at 4:50 PM Bruce Tate wrote:
> Maybe a transform function on with_index?
>
> Enum.with_index([:a, :b, :c], &{&2, &1})
>
> -bt
>
> On Tue, Dec 8, 2020 at 3:05 PM José Valim
Maybe a transform function on with_index?
Enum.with_index([:a, :b, :c], &{&2, &1})
-bt
On Tue, Dec 8, 2020 at 3:05 PM José Valim wrote:
> I would like to propose an Enum.index/2 function. It behaves like
> with_index/2 but the index is the first argument:
>
> iex> Enum.index([:a, :b, :c])
FWIW, I like the "indexed" function name; I think the past tense of the
verb helps indicate what it returns. The ambiguous nature of the verb/noun
"index" in English makes for confusion, for sure. "What does Enum.indexed
do?" "It returns the Enum, indexed." "Ah."
...Paul
On Tue, Dec 8,
I commonly need to flip the keys/values when using Map.new, why not add an
option to Map.new? I don’t like this name, but this reads pretty nicely to
me
list
|> Enum.with_index()
|> Map.new(flip: true)
On Tue, Dec 8, 2020 at 4:35 PM Austin Ziegler wrote:
> I’m not opposed to the idea, but
I’m not opposed to the idea, but wouldn’t that map be easily done with:
Map.new(Enum.with_index(data), fn {v, k} -> {k, v} end)
The biggest advantage of Map.new(Enum.index(data)) is that the number of
tuple creations is halved.
-a
On Tue, Dec 8, 2020 at 4:10 PM José Valim wrote:
> The
The main reason for an index function is to make it easier to build a map:
Map.new Enum.index(data)
I agree the duplicity between noun and verb is undesirable though.
On Tue, Dec 8, 2020 at 9:52 PM Kelvin Raffael Stinghen <
kelvin.sting...@gmail.com> wrote:
> I can't see why not just `x |>
I can't see why not just `x |> Stream.with_index() |> Stream.map(fn {v, i}
-> {i, v} end)`... I mean, lots of proposals for similar functions got
denied because it would be easy to develop a lib that does it cause Elixir
is awesomely extensible, so why bothering add another variant? Do you have
If this is related to the mystery benchmark tweets, José, then I guess I'm in
favor of the proposal :-)
I do prefer Felipe's name of `Enum.indexed/1` though. While `index` can be a
verb, it's ambiguous with the noun. `indexed` is clearly a past tense verb.
-Greg Vaughn
> On Dec 8, 2020, at
I'm not sure I see the benefit of the new Enum.index/2 over
Enum.with_index/2 and Enum.map/2.
That is to say, why would the language need both? I will admit it makes
piping into Enum.into/2 potentially nicer.
--
You received this message because you are subscribed to the Google Groups
That's a really good point.
Maybe `indexed`? `with_index` sounds like perfect past for me, so this
would be similar. Still `index` is fine.
About the function itself, I find it useful and believe it should be on
Enum.
Em ter., 8 de dez. de 2020 às 17:14, José Valim
escreveu:
> I considered the
I considered the option but I prefer to avoid using options that change the
return type.
On Tue, Dec 8, 2020 at 9:11 PM Felipe Stival wrote:
> I'm not sure. Sounds a bit ambiguous, couldn't this be an option for a new
> Enum.with_index/3?
>
> Em ter., 8 de dez. de 2020 às 17:05, José Valim
>
I'm not sure. Sounds a bit ambiguous, couldn't this be an option for a new
Enum.with_index/3?
Em ter., 8 de dez. de 2020 às 17:05, José Valim
escreveu:
> I would like to propose an Enum.index/2 function. It behaves like
> with_index/2 but the index is the first argument:
>
> iex>
I would like to propose an Enum.index/2 function. It behaves like
with_index/2 but the index is the first argument:
iex> Enum.index([:a, :b, :c])
[{0, :a}, {1, :b}, {2, :c}]
I am hoping it reads "index" this tuple (hence the index as the first
element. While with_index means attach an
It is not very easy to write multiline in doctests manually.
It would be great if autoformatter is applied to doctests too.
--
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,
20 matches
Mail list logo
