为什么 __next__() 内部的 yield 返回生成器 object？

Question

I am using yield to return the next value in the __next__() function in my class. However it does not return the next value, it returns the generator object.

I am trying to better understand iterators and yield . I might be doing it in the wrong way.

Have a look.

class MyString:
    def __init__(self,s):
        self.s=s

    def __iter__(self):
        return self

    def __next__(self):
        for i in range(len(self.s)):
            yield(self.s[i])

r=MyString("abc")
i=iter(r)
print(next(i))

This returns:

generator object __next__ at 0x032C05A0

Answer 1

next pretty much just calls __next__() in this case. Calling __next__ on your object will start the generator and return it (no magic is done at this point).

---

In this case, you might be able to get away with not defining __next__ at all:

class MyString:
    def __init__(self,s):
        self.s=s

    def __iter__(self):
        for i in range(len(self.s)):
            yield(self.s[i])
        # Or...
        # for item in self.s:
        #     yield item

If you wanted to use __iter__ and __next__ (to define an iterator rather than simply making an iterable ), you'd probably want to do something like this:

class MyString:
    def __init__(self,s):
        self.s = s
        self._ix = None

    def __iter__(self):
        return self

    def __next__(self):
        if self._ix is None:
            self._ix = 0

        try:
            item = self.s[self._ix]
        except IndexError:
            # Possibly reset `self._ix`?
            raise StopIteration
        self._ix += 1
        return item

Answer 2

Let's take a look at the purpose of the __next__ method. From the docs :

iterator.__next__()

Return the next item from the container. If there are no further items, raise the StopIteration exception.

Now let's see what the yield statement does. Another excerpt from the docs :

Using a yield expression in a function's body causes that function to be a generator

And

When a generator function is called, it returns an iterator known as a generator.

Now compare __next__ and yield : __next__ returns the next item from the container . But a function containing the yield keyword returns an iterator . Consequently, using yield in a __next__ method results in an iterator that yields iterators.

---

If you want to use yield to make your class iterable, do it in the __iter__ method:

class MyString:
    def __init__(self, s):
        self.s = s

    def __iter__(self):
        for s in self.s:
            yield s

The __iter__ method is supposed to return an iterator - and the yield keyword makes it do exactly that.

---

For completeness, here is how you would implement an iterator with a __next__ method. You have to keep track of the state of the iteration, and return the corresponding value. The easiest solution is probably to increment an index every time __next__ is called:

class MyString:
    def __init__(self,s):
        self.s = s
        self.index = -1

    def __iter__(self):
        return self

    def __next__(self):
        self.index += 1

        if self.index >= len(self.s):
            raise StopIteration

        return self.s[self.index]

Answer 3

As far as I can tell, generator functions are just syntactic sugar for classes with a next function. Example:

>>> def f():
    i = 0
    while True:
        i += 1
        yield i


>>> x = f()
>>> x
<generator object f at 0x0000000000659938>
>>> next(x)
1
>>> next(x)
2
>>> next(x)
3
>>> class g(object):
    def __init__(self):
        self.i = 0

    def __next__(self):
        self.i += 1
        return self.i


>>> y = g()
>>> y
<__main__.g object at 0x000000000345D908>
>>> next(y)
1
>>> next(y)
2
>>> next(y)
3

In fact, I came here looking to see if there is any significant difference. Please shout if there is.

So, to answer the question, what you have is a class with a __next__ method that returns an object that also has a __next__ method. So the simplest thing to do would be to replace your yield with a return and to keep track of how far along you are, and to remember to raise a StopIteration when you reach the end of the array. So something like:

class MyString:
    def __init__(self,s):
        self.s=s
        self._i = -1

    def __iter__(self):
        return self

    def __next__(self):
        self._i += 1
        if self._i >= len(self.s):
            raise StopIteration
        return self.s[self._i]

That's probably the simplest way to achieve what I think you're looking for.

Answer 4

OBSERVATION

If next() function calls __next__() method, what's happening in the following example.

Code:

class T:
    def __init__(self):
        self.s = 10
        
    def __iter__(self):
        for i in range(self.s):
            yield i
    
    def __next__(self):
        print('__next__ method is called.')
    
if __name__== '__main__':
    obj = T()
    k = iter(obj)
    print(next(k)) #0
    print(next(k)) #1
    print(next(k)) #2
    print(next(k)) #3
    print(next(k)) #4
    print(next(k)) #5
    print(next(k)) #6
    print(next(k)) #7
    print(next(k)) #8
    print(next(k)) #9
    print(next(k))
    print(next(k))

Terminal:

C:...>python test.py
0
1
2
3
4
5
6
7
8
9
Traceback (most recent call last):
  File "test.py", line 25, in <module>
    print(next(k))
StopIteration

WHAT IS HAPPENING?

It seams that next() function does not calling __next__ method. I cannot understand why python docs states that " next(iterator, default) Retrieve the next item from the iterator by calling its __next__() method." If someonw knows, let us help!

Case: __iter__ with __next__ in custom class with yield

So, if you want to use yield (in order to create a generator) with __iter__ and __next__ methods in a custom class , do not put just the yield into the __next__ method, but use it with __iter__(self) and return self.__next__() instead return self .

Code:

class T:
    def __init__(self):
        self.s = 10
        
    def __iter__(self):
        return self.__next__()
    
    def __next__(self):
        for i in range(self.s):
            yield i
    
if __name__== '__main__':
    obj = T()
    for i in obj:
        print(i)

Terminal:

C:\...>python test.py
0
1
2
3
4
5
6
7
8
9

C:...>

Also, you can call from __iter__ any other method instead __next__() .

Code:

class T:
    def __init__(self):
        self.s = 10
        
    def __iter__(self):
        return self.foo()
    
    def foo(self):
        for i in range(self.s):
            yield i
    
if __name__== '__main__':
    obj = T()
    for i in obj:
        print(i)

You will have exactly the same results.

Case: yield in __iter__ method without __next__ method

I don't think it is a good idea to use yield in __iter__ . Ok, it works, but I think that destroys the class API.

Case: __iter__ with __next__ in custom class without yield

Use these methods ( __iter__ and __next__ ). In the __iter__ return self and do not forget to raise StopIteration in __next__ method.

Code:

class T:
    def __init__(self):
        self.s = 10
        
    def __iter__(self):
        self.__i = -1
        return self
    
    def __next__(self):
        while  self.__i < self.s-1:
            self.__i+=1
            return self.__i
        raise StopIteration
    
if __name__== '__main__':
    obj = T()
    for i in obj:
        print(i)