Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.python > #76209 > unrolled thread

Why does str not have a __radd__ method?

Back to article view | Back to comp.lang.python

Contents

  Why does str not have a __radd__ method? Steven D'Aprano <steve+comp.lang.python@pearwood.info> - 2014-08-14 02:00 +1000
    Re: Why does str not have a __radd__ method? Ethan Furman <ethan@stoneleaf.us> - 2014-08-13 10:27 -0700
      Re: Why does str not have a __radd__ method? Steven D'Aprano <steve+comp.lang.python@pearwood.info> - 2014-08-14 03:55 +1000
        Re: Why does str not have a __radd__ method? Ethan Furman <ethan@stoneleaf.us> - 2014-08-13 11:22 -0700
        Re: Why does str not have a __radd__ method? Jason Swails <jason.swails@gmail.com> - 2014-08-13 14:36 -0400
          Re: Why does str not have a __radd__ method? wxjmfauth@gmail.com - 2014-08-13 23:24 -0700

#76209 — Why does str not have a __radd__ method?

I just tried to override str.__radd__:

class Special(str):
    def __radd__(self, other):
        print("I'm special!")
        return super().__radd__(self, other)

My __radd__ method was called correctly by the + operator, but to my
surprise, the super().__radd__ call failed with:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "<stdin>", line 4, in __radd__
AttributeError: 'super' object has no attribute '__radd__'

Sure enough, in both Python 3.3 and 2.7:

py> str.__radd__
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
AttributeError: type object 'str' has no attribute '__radd__'


This is especially astonishing, since int and float both have __radd__
methods, and yet numeric addition is commutative (x+y == y+x) whereas the
same is not true for string concatenation.

What is the rationale for str not having __radd__ method?


-- 
Steven

[toc] | [next] | [standalone]

#76219

On 08/13/2014 09:00 AM, Steven D'Aprano wrote:
>
> What is the rationale for str not having __radd__ method?

At a guess I would say because string only knows how to add itself to other strings, so __add__ is sufficient.

--
~Ethan~

[toc] | [prev] | [next] | [standalone]

#76221

Ethan Furman wrote:

> On 08/13/2014 09:00 AM, Steven D'Aprano wrote:
>>
>> What is the rationale for str not having __radd__ method?
> 
> At a guess I would say because string only knows how to add itself to
> other strings, so __add__ is sufficient.

# Python 2.7
py> "Hello" + u"World"
u'HelloWorld'
py> unicode.__radd__
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
AttributeError: type object 'unicode' has no attribute '__radd__'


My brain hurts.



-- 
Steven

[toc] | [prev] | [next] | [standalone]

#76224

On 08/13/2014 10:55 AM, Steven D'Aprano wrote:
> Ethan Furman wrote:
>
>> On 08/13/2014 09:00 AM, Steven D'Aprano wrote:
>>>
>>> What is the rationale for str not having __radd__ method?
>>
>> At a guess I would say because string only knows how to add itself to
>> other strings, so __add__ is sufficient.
>
> # Python 2.7
> py> "Hello" + u"World"
> u'HelloWorld'
> py> unicode.__radd__
> Traceback (most recent call last):
>    File "<stdin>", line 1, in <module>
> AttributeError: type object 'unicode' has no attribute '__radd__'

Well, unicode is a string type, right? ;)

And for the proof:

>>> 'hello'.__add__(u'world')
u'helloworld'
>>> u'hello'.__add__('world')
u'helloworld'


> My brain hurts.

An occupational hazard of unicode, surely.

--
~Ethan~

[toc] | [prev] | [next] | [standalone]

#76225

[Multipart message — attachments visible in raw view] — view raw

On Wed, Aug 13, 2014 at 1:55 PM, Steven D'Aprano <
steve+comp.lang.python@pearwood.info> wrote:

> Ethan Furman wrote:
>
> > On 08/13/2014 09:00 AM, Steven D'Aprano wrote:
> >>
> >> What is the rationale for str not having __radd__ method?
> >
> > At a guess I would say because string only knows how to add itself to
> > other strings, so __add__ is sufficient.
>
> # Python 2.7
> py> "Hello" + u"World"
> u'HelloWorld'
> py> unicode.__radd__
> Traceback (most recent call last):
>   File "<stdin>", line 1, in <module>
> AttributeError: type object 'unicode' has no attribute '__radd__'
>

This happens because the str.__add__ function calls string_concat under the
hood (see Objects/stringobject.c) -- there's a unicode check on the other
operand that results in the result of PyUnicode_Concat being returned
instead of the concatenated str type.  This doesn't require that unicode
define __radd__.

When the left-hand operand is Unicode, PyUnicode_Concat is called directly
(which is why the exception message is different for u'this' + 1 and 'this'
+ 1):

>>> 'this' + 1
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: cannot concatenate 'str' and 'int' objects
>>> u'this' + 1
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: coercing to Unicode: need string or buffer, int found


All the best,
Jason

[toc] | [prev] | [next] | [standalone]

#76266

Le mercredi 13 août 2014 20:36:55 UTC+2, Jason Swails a écrit :
> On Wed, Aug 13, 2014 at 1:55 PM, Steven D'Aprano <steve+comp....@pearwood.info> wrote:
> 
> 
> Ethan Furman wrote:
> 
> 
> 
> > On 08/13/2014 09:00 AM, Steven D'Aprano wrote:
> 
> >>
> 
> 
> >> What is the rationale for str not having __radd__ method?
> 
> >
> 
> 
> > At a guess I would say because string only knows how to add itself to
> 
> > other strings, so __add__ is sufficient.
> 
> 
> 
> # Python 2.7
> 
> py> "Hello" + u"World"
> 
> u'HelloWorld'
> 
> py> unicode.__radd__
> 
> 
> Traceback (most recent call last):
> 
>   File "<stdin>", line 1, in <module>
> 
> AttributeError: type object 'unicode' has no attribute '__radd__'
> 
> 
> 
> This happens because the str.__add__ function calls string_concat under the hood (see Objects/stringobject.c) -- there's a unicode check on the other operand that results in the result of PyUnicode_Concat being returned instead of the concatenated str type.  This doesn't require that unicode define __radd__.
> 
> 
> 
> When the left-hand operand is Unicode, PyUnicode_Concat is called directly (which is why the exception message is different for u'this' + 1 and 'this' + 1):
> 
> 
> 
> 
> >>> 'this' + 1
> Traceback (most recent call last):
> 
>   File "<stdin>", line 1, in <module>
> TypeError: cannot concatenate 'str' and 'int' objects
> 
> >>> u'this' + 1
> Traceback (most recent call last):
>   File "<stdin>", line 1, in <module>
> 
> TypeError: coercing to Unicode: need string or buffer, int found
> 
> 
> 
> 
> All the best,
> Jason

--------

This is more funny.

>>> print u'a' + 'a'
aa
>>> print 'a' + u'a'
aa
>>> print u'a' + 'é'

Traceback (most recent call last):
  File "<pyshell#4>", line 1, in <module>
    print u'a' + 'é'
UnicodeDecodeError: 'ascii' codec can't decode byte 0xe9 in position 0: ordinal not in range(128)
>>> 

Python? The single language, which tried (and still tries
to make the ascii world compatible with the unicode
world.

In a somwhow symmetrical way, but a little bit
differently, the situation in Python 3 is even worse.

jmf

PS See my comment about the Euro sign in a previous
thread.

[toc] | [prev] | [standalone]

Back to top | Article view | comp.lang.python

csiph-web