[Python-ideas] Adding "+" and "+=" operators to dict

[Nikolaus Rath]

On Feb 13 2015, Erik Bray <erik.m.bray-Re5JQEeQqe8AvxtiuMwx3w at public.gmane.org> wrote:
> On Fri, Feb 13, 2015 at 6:22 PM, Nikolaus Rath <Nikolaus-BTH8mxji4b0 at public.gmane.org> wrote:
>> Alexander Heger <python-ePO413wvQzY-XMD5yJDbdMReXY1tMh2IBg at public.gmane.org> writes:
>>>> As far as I was, and still am, concerned, & is the obvious and most
>>>> natural operator for concatenation. [1, 2]+[3, 4] should return [4, 6],
>>>> and sum(bunch of lists) should be a meaningless operation, like
>>>> sum(bunch of HTTP servers). Or sum(bunch of dicts).
>>>
>>> This only works if the items *can* be added.  Dics should not make
>>> such an assumption. & is not a more natural operator, because, why
>>> would you then not just expect that [1, 2] & [3, 4] returns [1 & 2, 3
>>> & 4] == [0 , 0] ?  the same would be true for any operator you pick.
>>
>> Which brings us back to the idea to introduce elementwise variants of
>> any operator:
>>
>> [1,2] .+ [3,4] == [1+3, 2+4]
>> [1,2] + [3,4] == [1,2,3,4]
>> [1,2] .& [3,4] == [1 & 3, 2 & 4]
>> [1,2] & [3,4] == not (yet?) defined
>>
>> As a regular numpy user, I'd be very happy about that too :-).
>
> ... except for the part where in Numpy ".+" and "+" and so one would

Parse error.

> have to be identical, which would be no end of confusing especially
> when adding, say, a Numpy array and a list.


Best,
-Nikolaus

-- 
GPG encrypted emails preferred. Key id: 0xD113FCAC3C4E599F
Fingerprint: ED31 791B 2C5C 1613 AF38 8B8A D113 FCAC 3C4E 599F

             »Time flies like an arrow, fruit flies like a Banana.«