I don't think this will turn the result of this discussion, but I've some
comments on:
> I personally can't recall a need for put_new!, especially because *the
best use of "static maps" is by defining all keys upfront and then
updating/replacing them*.
I've never needed that for "static maps", but I've wanted put_new! a few
times with dynamic maps for exactly the opposite of replace!/update!.
Consider a server receiving messages for CRUD operations, where entries are
identified by some id. On "create" I only want to add a record if no entry
for a given id exists (put_new! or first write wins for create) and on
"update" I want to make sure an entry for the given id exists to apply the
update to.
Mário Guimarães schrieb am Mittwoch, 31. Juli 2019 um 19:59:52 UTC+2:
> I have implemented like this
>
> defmodule KeyFoundError do
> defexception [:key, :term]
> def message(%__MODULE__{key: key, term: term}) do
> "key #{inspect(key)} found in: #{inspect(term)}"
> end
> end
>
>
> defmodule MyMap do
> def put_new!(map, key, value) do
> if Map.has_key?(map, key) do
> raise KeyFoundError, key: key, term: map
> else
> Map.put(map, key, value)
> end
> end
> end
>
> and use it to ensure that an algorithm building a map with a
> non-predefined set of keys fulfills the requirement to write each key only
> once, which if it is not the case may be a signal that something is wrong
> with the algorithm.
>
> I agree that this is perhaps only a need in more complex, and thus, less
> common algorithms, like in my use case, where the keys come from an
> external data source, and I want to be sure that this is not sending the
> same key duplicated.
>
> Mário
>
>
> segunda-feira, 29 de Julho de 2019 às 11:27:07 UTC+1, José Valim escreveu:
>>
>> > 1. put_new and put_new! pair behaves differently from !-paired
>> functions, I’m not sure if it’s good thing to break expectations in that
>> case.
>>
>> I believe the concern here is that while all functions raise when the key
>> does not exist, the proposed put_new! raises when the key exists. It is how
>> put_new! should behave... but it is a different behaviour than the one that
>> already exists, highlighted by the fact a new exception would have to be
>> added.
>>
>> Ben commented on the second point already.
>>
>> Also, your proposal does not include any code examples or snippets on how
>> this pattern would be used. I personally can't recall a need for put_new!,
>> especially because *the best use of "static maps" is by defining all keys
>> upfront and then updating/replacing them*.
>>
>> *José Valim*
>> www.plataformatec.com.br
>> Skype: jv.ptec
>> Founder and Director of R&D
>>
>>
--
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 elixir-lang-core+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/elixir-lang-core/b13a362c-2028-4073-90f1-79216cb51067n%40googlegroups.com.
