If you want Unicode characters, then you can do:
for <<char::utf8 <- input>>, do: maybe_replace(char), into: ""
Note we are matching in characters though, so:
defp maybe_replace(?A), do: "a"
defp maybe_replace(?€), do: "E"
...
defp maybe_replace(char), do: <<char::utf8>>
It won't work for graphemes though (which is a much more complicated
problem because the same grapheme can be in different forms and their
interpretation may also be localy dependent).
*José Valim*
www.plataformatec.com.br
Skype: jv.ptec
Founder and Director of R&D
On Fri, Jun 28, 2019 at 2:08 PM Krep Flap <[email protected]> wrote:
> Thanks for your reply.
>
> I agree that adding syntactic sugar purely for an ultra-concise
> implementation is probably unneeded, given the options and flexibility you
> demonstrated.
>
> In your example, what if we also want to replace multi-byte characters,
> e.g. replace "€" with "E"?
>
> Since we cannot use the bitstring generator in that case, is it still
> efficient to use a charlist to get the codepoints?
>
> for char <- to_charlist(input), do: maybe_replace(char), into: ""
>
> defp maybe_replace(?A), do: "a"
> defp maybe_replace(?€), do: "E"
> ...
> defp maybe_replace(char), do: <<char>>
>
>
> Op vrijdag 28 juni 2019 10:56:09 UTC+2 schreef José Valim:
>>
>> Thanks for the proposal!
>>
>> I assume those languages do not have a powerful mechanism for traversing
>> binaries like Elixir does though?
>>
>> In Elixir you could do:
>>
>> for <<char <- input>>, do: maybe_replace(char), into: ""
>>
>> Note we are matching in characters though, so:
>>
>> defp maybe_replace(?A), do: "a"
>> ...
>> defp maybe_replace(char), do: <<char>>
>>
>> You can also use pattern matching and recursion.
>>
>> Both variants should be quite more efficient than an API that expects a
>> map and do map look-ups. It also gives you more control over what you are
>> matching (is it characters? Unicode codepoints? 16-bits integers?).
>>
>> But if for some reason you need map look-ups:
>>
>> map = %{
>> ?A => "a"
>> }
>> for <<char <- input>>, do: Map.get(map, char) || <<char>>, into: ""
>>
>> And finally, if you want to match on substrings (requires Elixir v1.9):
>>
>> map = %{
>> "AA" => "a"
>> }
>> String.replace(input, Map.keys(map), &map[&1])
>>
>> TL;DR - I believe we have more performant and more flexible ways of
>> achieving the same feature to not warrant the particular "tr"
>> implementation found in other languages.
>>
>> *José Valim*
>> www.plataformatec.com.br
>> Skype: jv.ptec
>> Founder and Director of R&D
>>
>>
>> On Fri, Jun 28, 2019 at 10:44 AM <[email protected]> wrote:
>>
>>> Recently I found myself wanting to replace certain characters in a
>>> string with others (similar to "tr" shell command).
>>>
>>> In my case I wanted to replace any of the characters in :;<=>?@[\]^_`
>>> with the corresponding character from ABCDEFGabcdef.
>>>
>>> I'm not sure what the most idiomatic way to write this would be in
>>> Elixir but I used:
>>>
>>> Enum.map_join(input, &my_replace/1)
>>>
>>> defp my_replace(":"), do: "A"
>>> defp my_replace(";"), do: "B"
>>> defp my_replace("<"), do: "C"
>>> defp my_replace("="), do: "D"
>>> defp my_replace(">"), do: "E"
>>> defp my_replace("?"), do: "F"
>>> defp my_replace("@"), do: "G"
>>> defp my_replace("["), do: "a"
>>> defp my_replace("\\"), do: "b"
>>> defp my_replace("]"), do: "c"
>>> defp my_replace("^"), do: "d"
>>> defp my_replace("_"), do: "e"
>>> defp my_replace("`"), do: "f"
>>> defp my_replace(char), do: char
>>>
>>> This works, is readable, but not concise. My proposal would thus be to
>>> open a discussion on the way to implement a more concise way to do this,
>>> unless there is already a concise way to do this that I'm not aware of and
>>> if people agree that it's worth implementing.
>>>
>>> This has been partly discussed before in
>>> https://github.com/elixir-lang/elixir/issues/4473
>>>
>>> Here are some reference implementations in other languages, to open up
>>> ideas:
>>>
>>> PHP:
>>>
>>> strtr($input, ":;<=>?@[\]^_`", "ABCDEFGabcdef")
>>>
>>> Ruby:
>>>
>>> input.tr!(':;<=>?@[\\]^_`', 'ABCDEFGabcdef')
>>>
>>> Go has an implementation that is not concise but does allow multiple
>>> char/string translation:
>>>
>>> r := strings.NewReplacer(
>>> ":", "A",
>>> ";", "B",
>>> "<", "C",
>>> // and so on...
>>> )
>>> r.Replace(input)
>>>
>>> Python (3) requires creation of a "translation map" first:
>>>
>>> table = str.maketrans(":;<=>?@[\]^_`", "ABCDEFGabcdef")
>>> input.translate(table)
>>>
>>> A lot of these implementations also allow other uses, besides simple
>>> character translations.
>>>
>>> --
>>> 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/LiSFMq0--3-1%40tuta.io
>>> <https://groups.google.com/d/msgid/elixir-lang-core/LiSFMq0--3-1%40tuta.io?utm_medium=email&utm_source=footer>
>>> .
>>> For more options, visit https://groups.google.com/d/optout.
>>>
>> --
> 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/2e11c739-9771-4623-94ba-5e9159bdc675%40googlegroups.com
> <https://groups.google.com/d/msgid/elixir-lang-core/2e11c739-9771-4623-94ba-5e9159bdc675%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
> For more options, visit https://groups.google.com/d/optout.
>
--
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/CAGnRm4L%2B2f1ihAOCEb-vY5%2BsjYc7%3DYsoex%3DQYa6%3DC4%3D6xyDppw%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.
