Why doesn't the Python interpreter implicitly create the generator?

Question

#def func(param):
#    if param < 0:
#        return "test"
#    i = 0
#    while i < param:
#        yield i
#        i += 1
def func(param):
    if param < 0:
        return "test"
    def gen(n):
        i = 0
        while i < param:
            yield i
            i += 1
    return gen(param)

print(func(-1))
print(func(3))
g = func(3)
for i in range(0, 3):
    print(next(g))

Is there a reason that the Python interpreter can not convert the commented code to the actual code implicitly? This seems like this should be allowed, but I am wondering what repercussions there are that made them choose to disallow this.

agf · Accepted Answer · 2013-07-24 04:47:37Z

2

When you call a generator function, you know what type of object it will return -- a generator. If we allowed generator functions to return, then you'd have to check the type of the return value before you could start iterating over the generator, so you'd no longer have a generator function -- just a function that might return a generator.

answered Jul 24, 2013 at 4:47

agf

178k45 gold badges300 silver badges241 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

zhangyangyu · Accepted Answer · 2013-07-24 05:14:22Z

2

In python2.x, you can not return something in a generator:

>>> def func():
...     return 3
...     yield 3
... 
  File "<stdin>", line 3
SyntaxError: 'return' with argument inside generator
>>>

In python3.x, use return in a generator means raise a StopIteration(<something>):

>>> def func():
...     return 3
...     yield 3
... 
>>> func().__next__()
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
StopIteration: 3
>>>

I can not think about any reason for the interpreter to decide which part is a generator. It is hard and I think it is the responsibilities of programmers. And I even doubt whether return a value in a generator is a good implementation.

edited Jul 24, 2013 at 5:14

answered Jul 24, 2013 at 4:57

zhangyangyu

8,6103 gold badges35 silver badges43 bronze badges

4 Comments

user2357112 Over a year ago

I thought that feature had gone in, but I could never find the source again after the first time I read it, and I don't have a Python 3 install to test with. Do you have a link to a PEP or something documenting the change?

zhangyangyu Over a year ago

PEP380. You can see the rule.@user2357112

Amber Over a year ago

Technically, you can use return in a generator in Python 2.x - you just can't return a value. It's equivalent to raising StopIteration.

agf Over a year ago

I'm not sure this answers the question, though it's good information. You don't explain why Python doesn't allow you to return a value from a generator function.

dbr · Accepted Answer · 2013-07-24 05:14:55Z

The reason is simply, if the def contains a yield statement, it creates a generator:

http://www.python.org/dev/peps/pep-0255/

The yield statement may only be used inside functions. A function that contains a yield statement is called a generator function. A generator function is an ordinary function object in all respects, but has the new CO_GENERATOR flag set in the code object's co_flags member.

That is how the interpreter distinguishes between a regular function, and a generator function. It's simple to implement, easy to reason about ("if it contains a yield, it's a generator")

The "conditional generator" behaviour you describe would be much more complex to implement, and in some cases not desirable (maybe the conditional should happen inside the first iteration of the generator, or maybe it should run as soon as you call func(...))

Your other code either returns a generator, or a string. If that's the interface you want, it seems like a perfectly good solution (but it's hard to make practical suggestions without a real example)

Aviv Bar-el · Accepted Answer · 2013-07-24 04:56:05Z

0

generators are declared like functions but are different, you cant use a return statement in generator what you should do is:

def func(param):
    if param < 0:
        yield "test"
        raise StopIteration()
    i = 0
    while i < param:
        yield i
        i += 1

an even better implmention would be:

def func(param):
    if param < 0:
        raise ValueError("param must be a positive number")
    i = 0
    while i < param:
        yield i
        i += 1

answered Jul 24, 2013 at 4:56

Aviv Bar-el

814 bronze badges

Collectives™ on Stack Overflow

Why doesn't the Python interpreter implicitly create the generator?

4 Answers 4

Comments

4 Comments

Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

4 Answers 4

Comments

4 Comments

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related