[python] What does the "yield" keyword do?

What is the use of the yield keyword in Python, and what does it do?

For example, I'm trying to understand this code¹:

def _get_child_candidates(self, distance, min_dist, max_dist):
    if self._leftchild and distance - max_dist < self._median:
        yield self._leftchild
    if self._rightchild and distance + max_dist >= self._median:
        yield self._rightchild  

And this is the caller:

result, candidates = [], [self]
while candidates:
    node = candidates.pop()
    distance = node._get_dist(obj)
    if distance <= max_dist and distance >= min_dist:
        result.extend(node._values)
    candidates.extend(node._get_child_candidates(distance, min_dist, max_dist))
return result

What happens when the method _get_child_candidates is called? Is a list returned? A single element? Is it called again? When will subsequent calls stop?

_{1. This piece of code was written by Jochen Schulz (jrschulz), who made a great Python library for metric spaces. This is the link to the complete source: Module mspace.}

Yet another TL;DR

Iterator on list: next() returns the next element of the list

Iterator generator: next() will compute the next element on the fly (execute code)

You can see the yield/generator as a way to manually run the control flow from outside (like continue loop one step), by calling next, however complex the flow.

Note: The generator is NOT a normal function. It remembers the previous state like local variables (stack). See other answers or articles for detailed explanation. The generator can only be iterated on once. You could do without yield, but it would not be as nice, so it can be considered 'very nice' language sugar.

The yield keyword is going to replace return in a function definition to create a generator.

def create_generator():
   for i in range(100):
   yield i
myGenerator = create_generator()
print(myGenerator)
# <generator object create_generator at 0x102dd2480>
for i in myGenerator:
   print(i) # prints 0-99

When the returned generator is first used—not in the assignment but the for loop—the function definition will execute until it reaches the yield statement. There, it will pause (see why it’s called yield) until used again. Then, it will pick up where it left off. Upon the final iteration of the generator, any code after the yield command will execute.

def create_generator():
   print("Beginning of generator")
   for i in range(4):
      yield i
   print("After yield")
print("Before assignment")

myGenerator = create_generator()

print("After assignment")
for i in myGenerator :
   print(i) # prints 0-3
"""
Before assignment
After assignment
Beginning of generator
0
1
2
After yield

The yield keyword modifies a function’s behavior to produce a generator that’s paused at each yield command during iteration. The function isn’t executed except upon iteration, which leads to improved resource management, and subsequently, a better overall performance. Use generators (and yielded functions) for creating large data sets meant for single-use iteration.

While a lot of answers show why you'd use a yield to create a generator, there are more uses for yield. It's quite easy to make a coroutine, which enables the passing of information between two blocks of code. I won't repeat any of the fine examples that have already been given about using yield to create a generator.

To help understand what a yield does in the following code, you can use your finger to trace the cycle through any code that has a yield. Every time your finger hits the yield, you have to wait for a next or a send to be entered. When a next is called, you trace through the code until you hit the yield… the code on the right of the yield is evaluated and returned to the caller… then you wait. When next is called again, you perform another loop through the code. However, you'll note that in a coroutine, yield can also be used with a send… which will send a value from the caller into the yielding function. If a send is given, then yield receives the value sent, and spits it out the left hand side… then the trace through the code progresses until you hit the yield again (returning the value at the end, as if next was called).

For example:

>>> def coroutine():
...     i = -1
...     while True:
...         i += 1
...         val = (yield i)
...         print("Received %s" % val)
...
>>> sequence = coroutine()
>>> sequence.next()
0
>>> sequence.next()
Received None
1
>>> sequence.send('hello')
Received hello
2
>>> sequence.close()

There's one extra thing to mention: a function that yields doesn't actually have to terminate. I've written code like this:

def fib():
    last, cur = 0, 1
    while True: 
        yield cur
        last, cur = cur, last + cur

Then I can use it in other code like this:

for f in fib():
    if some_condition: break
    coolfuncs(f);

It really helps simplify some problems, and makes some things easier to work with.

The yield keyword is reduced to two simple facts:

1. If the compiler detects the yield keyword anywhere inside a function, that function no longer returns via the return statement. Instead, it immediately returns a lazy "pending list" object called a generator
2. A generator is iterable. What is an iterable? It's anything like a list or set or range or dict-view, with a built-in protocol for visiting each element in a certain order.

In a nutshell: a generator is a lazy, incrementally-pending list, and yield statements allow you to use function notation to program the list values the generator should incrementally spit out.

generator = myYieldingFunction(...)
x = list(generator)

   generator
       v
[x[0], ..., ???]

         generator
             v
[x[0], x[1], ..., ???]

               generator
                   v
[x[0], x[1], x[2], ..., ???]

                       StopIteration exception
[x[0], x[1], x[2]]     done

list==[x[0], x[1], x[2]]

Example

Let's define a function makeRange that's just like Python's range. Calling makeRange(n) RETURNS A GENERATOR:

def makeRange(n):
    # return 0,1,2,...,n-1
    i = 0
    while i < n:
        yield i
        i += 1

>>> makeRange(5)
<generator object makeRange at 0x19e4aa0>

To force the generator to immediately return its pending values, you can pass it into list() (just like you could any iterable):

>>> list(makeRange(5))
[0, 1, 2, 3, 4]

Comparing example to "just returning a list"

The above example can be thought of as merely creating a list which you append to and return:

# list-version                   #  # generator-version
def makeRange(n):                #  def makeRange(n):
    """return [0,1,2,...,n-1]""" #~     """return 0,1,2,...,n-1"""
    TO_RETURN = []               #>
    i = 0                        #      i = 0
    while i < n:                 #      while i < n:
        TO_RETURN += [i]         #~         yield i
        i += 1                   #          i += 1  ## indented
    return TO_RETURN             #>

>>> makeRange(5)
[0, 1, 2, 3, 4]

There is one major difference, though; see the last section.

How you might use generators

An iterable is the last part of a list comprehension, and all generators are iterable, so they're often used like so:

#                   _ITERABLE_
>>> [x+10 for x in makeRange(5)]
[10, 11, 12, 13, 14]

To get a better feel for generators, you can play around with the itertools module (be sure to use chain.from_iterable rather than chain when warranted). For example, you might even use generators to implement infinitely-long lazy lists like itertools.count(). You could implement your own def enumerate(iterable): zip(count(), iterable), or alternatively do so with the yield keyword in a while-loop.

Please note: generators can actually be used for many more things, such as implementing coroutines or non-deterministic programming or other elegant things. However, the "lazy lists" viewpoint I present here is the most common use you will find.

Behind the scenes

This is how the "Python iteration protocol" works. That is, what is going on when you do list(makeRange(5)). This is what I describe earlier as a "lazy, incremental list".

>>> x=iter(range(5))
>>> next(x)
0
>>> next(x)
1
>>> next(x)
2
>>> next(x)
3
>>> next(x)
4
>>> next(x)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
StopIteration

The built-in function next() just calls the objects .next() function, which is a part of the "iteration protocol" and is found on all iterators. You can manually use the next() function (and other parts of the iteration protocol) to implement fancy things, usually at the expense of readability, so try to avoid doing that...

Minutiae

Normally, most people would not care about the following distinctions and probably want to stop reading here.

In Python-speak, an iterable is any object which "understands the concept of a for-loop" like a list [1,2,3], and an iterator is a specific instance of the requested for-loop like [1,2,3].__iter__(). A generator is exactly the same as any iterator, except for the way it was written (with function syntax).

When you request an iterator from a list, it creates a new iterator. However, when you request an iterator from an iterator (which you would rarely do), it just gives you a copy of itself.

Thus, in the unlikely event that you are failing to do something like this...

> x = myRange(5)
> list(x)
[0, 1, 2, 3, 4]
> list(x)
[]

... then remember that a generator is an iterator; that is, it is one-time-use. If you want to reuse it, you should call myRange(...) again. If you need to use the result twice, convert the result to a list and store it in a variable x = list(myRange(5)). Those who absolutely need to clone a generator (for example, who are doing terrifyingly hackish metaprogramming) can use itertools.tee if absolutely necessary, since the copyable iterator Python PEP standards proposal has been deferred.

I was going to post "read page 19 of Beazley's 'Python: Essential Reference' for a quick description of generators", but so many others have posted good descriptions already.

Also, note that yield can be used in coroutines as the dual of their use in generator functions. Although it isn't the same use as your code snippet, (yield) can be used as an expression in a function. When a caller sends a value to the method using the send() method, then the coroutine will execute until the next (yield) statement is encountered.

Generators and coroutines are a cool way to set up data-flow type applications. I thought it would be worthwhile knowing about the other use of the yield statement in functions.

Yet another TL;DR

Iterator on list: next() returns the next element of the list

Iterator generator: next() will compute the next element on the fly (execute code)

You can see the yield/generator as a way to manually run the control flow from outside (like continue loop one step), by calling next, however complex the flow.

Note: The generator is NOT a normal function. It remembers the previous state like local variables (stack). See other answers or articles for detailed explanation. The generator can only be iterated on once. You could do without yield, but it would not be as nice, so it can be considered 'very nice' language sugar.

Here is an example in plain language. I will provide a correspondence between high-level human concepts to low-level Python concepts.

I want to operate on a sequence of numbers, but I don't want to bother my self with the creation of that sequence, I want only to focus on the operation I want to do. So, I do the following:

• I call you and tell you that I want a sequence of numbers which is produced in a specific way, and I let you know what the algorithm is.
  This step corresponds to defining the generator function, i.e. the function containing a yield.
• Sometime later, I tell you, "OK, get ready to tell me the sequence of numbers".
  This step corresponds to calling the generator function which returns a generator object. Note that you don't tell me any numbers yet; you just grab your paper and pencil.
• I ask you, "tell me the next number", and you tell me the first number; after that, you wait for me to ask you for the next number. It's your job to remember where you were, what numbers you have already said, and what is the next number. I don't care about the details.
  This step corresponds to calling .next() on the generator object.
• … repeat previous step, until…
• eventually, you might come to an end. You don't tell me a number; you just shout, "hold your horses! I'm done! No more numbers!"
  This step corresponds to the generator object ending its job, and raising a StopIteration exception The generator function does not need to raise the exception. It's raised automatically when the function ends or issues a return.

This is what a generator does (a function that contains a yield); it starts executing, pauses whenever it does a yield, and when asked for a .next() value it continues from the point it was last. It fits perfectly by design with the iterator protocol of Python, which describes how to sequentially request values.

The most famous user of the iterator protocol is the for command in Python. So, whenever you do a:

for item in sequence:

it doesn't matter if sequence is a list, a string, a dictionary or a generator object like described above; the result is the same: you read items off a sequence one by one.

Note that defining a function which contains a yield keyword is not the only way to create a generator; it's just the easiest way to create one.

For more accurate information, read about iterator types, the yield statement and generators in the Python documentation.

All great answers, however a bit difficult for newbies.

I assume you have learned the return statement.

As an analogy, return and yield are twins. return means 'return and stop' whereas 'yield` means 'return, but continue'

1. Try to get a num_list with return.

def num_list(n):
    for i in range(n):
        return i

Run it:

In [5]: num_list(3)
Out[5]: 0

See, you get only a single number rather than a list of them. return never allows you prevail happily, just implements once and quit.

2. There comes yield

Replace return with yield:

In [10]: def num_list(n):
    ...:     for i in range(n):
    ...:         yield i
    ...:

In [11]: num_list(3)
Out[11]: <generator object num_list at 0x10327c990>

In [12]: list(num_list(3))
Out[12]: [0, 1, 2]

Now, you win to get all the numbers.

Comparing to return which runs once and stops, yield runs times you planed. You can interpret return as return one of them, and yield as return all of them. This is called iterable.

3. One more step we can rewrite yield statement with return

In [15]: def num_list(n):
    ...:     result = []
    ...:     for i in range(n):
    ...:         result.append(i)
    ...:     return result

In [16]: num_list(3)
Out[16]: [0, 1, 2]

It's the core about yield.

The difference between a list return outputs and the object yield output is:

You will always get [0, 1, 2] from a list object but only could retrieve them from 'the object yield output' once. So, it has a new name generator object as displayed in Out[11]: <generator object num_list at 0x10327c990>.

In conclusion, as a metaphor to grok it:

• return and yield are twins
• list and generator are twins

First yield program

def countdown_gen(x):
      count = x
      while count > 0:
           yield count
           count -= 1

g = countdown_gen(5)

for item in g:
     print(item)

Output

5
4
3
2
1

Understand flow

Note

1. Return a generator obj Suspend a functon execution and save its
2. status, function can be executed again.

In Python generators (a special type of iterators) are used to generate series of values and yield keyword is just like the return keyword of generator functions.

The other fascinating thing yield keyword does is saving the state of a generator function.

So, we can set a number to a different value each time the generator yields.

Here's an instance:

def getPrimes(number):
    while True:
        if isPrime(number):
            number = yield number     # a miracle occurs here
        number += 1

def printSuccessivePrimes(iterations, base=10):
    primeGenerator = getPrimes(base)
    primeGenerator.send(None)
    for power in range(iterations):
        print(primeGenerator.send(base ** power))

It's returning a generator. I'm not particularly familiar with Python, but I believe it's the same kind of thing as C#'s iterator blocks if you're familiar with those.

The key idea is that the compiler/interpreter/whatever does some trickery so that as far as the caller is concerned, they can keep calling next() and it will keep returning values - as if the generator method was paused. Now obviously you can't really "pause" a method, so the compiler builds a state machine for you to remember where you currently are and what the local variables etc look like. This is much easier than writing an iterator yourself.

Here is an example in plain language. I will provide a correspondence between high-level human concepts to low-level Python concepts.

I want to operate on a sequence of numbers, but I don't want to bother my self with the creation of that sequence, I want only to focus on the operation I want to do. So, I do the following:

• I call you and tell you that I want a sequence of numbers which is produced in a specific way, and I let you know what the algorithm is.
  This step corresponds to defining the generator function, i.e. the function containing a yield.
• Sometime later, I tell you, "OK, get ready to tell me the sequence of numbers".
  This step corresponds to calling the generator function which returns a generator object. Note that you don't tell me any numbers yet; you just grab your paper and pencil.
• I ask you, "tell me the next number", and you tell me the first number; after that, you wait for me to ask you for the next number. It's your job to remember where you were, what numbers you have already said, and what is the next number. I don't care about the details.
  This step corresponds to calling .next() on the generator object.
• … repeat previous step, until…
• eventually, you might come to an end. You don't tell me a number; you just shout, "hold your horses! I'm done! No more numbers!"
  This step corresponds to the generator object ending its job, and raising a StopIteration exception The generator function does not need to raise the exception. It's raised automatically when the function ends or issues a return.

This is what a generator does (a function that contains a yield); it starts executing, pauses whenever it does a yield, and when asked for a .next() value it continues from the point it was last. It fits perfectly by design with the iterator protocol of Python, which describes how to sequentially request values.

The most famous user of the iterator protocol is the for command in Python. So, whenever you do a:

for item in sequence:

it doesn't matter if sequence is a list, a string, a dictionary or a generator object like described above; the result is the same: you read items off a sequence one by one.

Note that defining a function which contains a yield keyword is not the only way to create a generator; it's just the easiest way to create one.

For more accurate information, read about iterator types, the yield statement and generators in the Python documentation.

Yield gives you a generator.

def get_odd_numbers(i):
    return range(1, i, 2)
def yield_odd_numbers(i):
    for x in range(1, i, 2):
       yield x
foo = get_odd_numbers(10)
bar = yield_odd_numbers(10)
foo
[1, 3, 5, 7, 9]
bar
<generator object yield_odd_numbers at 0x1029c6f50>
bar.next()
1
bar.next()
3
bar.next()
5

As you can see, in the first case foo holds the entire list in memory at once. It's not a big deal for a list with 5 elements, but what if you want a list of 5 million? Not only is this a huge memory eater, it also costs a lot of time to build at the time that the function is called.

In the second case, bar just gives you a generator. A generator is an iterable--which means you can use it in a for loop, etc, but each value can only be accessed once. All the values are also not stored in memory at the same time; the generator object "remembers" where it was in the looping the last time you called it--this way, if you're using an iterable to (say) count to 50 billion, you don't have to count to 50 billion all at once and store the 50 billion numbers to count through.

Again, this is a pretty contrived example, you probably would use itertools if you really wanted to count to 50 billion. :)

This is the most simple use case of generators. As you said, it can be used to write efficient permutations, using yield to push things up through the call stack instead of using some sort of stack variable. Generators can also be used for specialized tree traversal, and all manner of other things.

It's returning a generator. I'm not particularly familiar with Python, but I believe it's the same kind of thing as C#'s iterator blocks if you're familiar with those.

The key idea is that the compiler/interpreter/whatever does some trickery so that as far as the caller is concerned, they can keep calling next() and it will keep returning values - as if the generator method was paused. Now obviously you can't really "pause" a method, so the compiler builds a state machine for you to remember where you currently are and what the local variables etc look like. This is much easier than writing an iterator yourself.

yield in python is in a way similar to the return statement, except for some differences. If multiple values have to be returned from a function, return statement will return all the values as a list and it has to be stored in the memory in the caller block. But what if we don't want to use extra memory? Instead, we want to get the value from the function when we need it. This is where yield comes in. Consider the following function :-

def fun():
   yield 1
   yield 2
   yield 3

And the caller is :-

def caller():
   print ('First value printing')
   print (fun())
   print ('Second value printing')
   print (fun())
   print ('Third value printing')
   print (fun())

The above code segment (caller function) when called, outputs :-

First value printing
1
Second value printing
2
Third value printing
3

As can be seen from above, yield returns a value to its caller, but when the function is called again, it doesn't start from the first statement, but from the statement right after the yield. In the above example, "First value printing" was printed and the function was called. 1 was returned and printed. Then "Second value printing" was printed and again fun() was called. Instead of printing 1 (the first statement), it returned 2, i.e., the statement just after yield 1. The same process is repeated further.

In summary, the yield statement transforms your function into a factory that produces a special object called a generator which wraps around the body of your original function. When the generator is iterated, it executes your function until it reaches the next yield then suspends execution and evaluates to the value passed to yield. It repeats this process on each iteration until the path of execution exits the function. For instance,

def simple_generator():
    yield 'one'
    yield 'two'
    yield 'three'

for i in simple_generator():
    print i

simply outputs

one
two
three

The power comes from using the generator with a loop that calculates a sequence, the generator executes the loop stopping each time to 'yield' the next result of the calculation, in this way it calculates a list on the fly, the benefit being the memory saved for especially large calculations

Say you wanted to create a your own range function that produces an iterable range of numbers, you could do it like so,

def myRangeNaive(i):
    n = 0
    range = []
    while n < i:
        range.append(n)
        n = n + 1
    return range

and use it like this;

for i in myRangeNaive(10):
    print i

But this is inefficient because

• You create an array that you only use once (this wastes memory)
• This code actually loops over that array twice! :(

Luckily Guido and his team were generous enough to develop generators so we could just do this;

def myRangeSmart(i):
    n = 0
    while n < i:
       yield n
       n = n + 1
    return

for i in myRangeSmart(10):
    print i

Now upon each iteration a function on the generator called next() executes the function until it either reaches a 'yield' statement in which it stops and 'yields' the value or reaches the end of the function. In this case on the first call, next() executes up to the yield statement and yield 'n', on the next call it will execute the increment statement, jump back to the 'while', evaluate it, and if true, it will stop and yield 'n' again, it will continue that way until the while condition returns false and the generator jumps to the end of the function.

Many people use return rather than yield, but in some cases yield can be more efficient and easier to work with.

Here is an example which yield is definitely best for:

return (in function)

import random

def return_dates():
    dates = [] # With 'return' you need to create a list then return it
    for i in range(5):
        date = random.choice(["1st", "2nd", "3rd", "4th", "5th", "6th", "7th", "8th", "9th", "10th"])
        dates.append(date)
    return dates

yield (in function)

def yield_dates():
    for i in range(5):
        date = random.choice(["1st", "2nd", "3rd", "4th", "5th", "6th", "7th", "8th", "9th", "10th"])
        yield date # 'yield' makes a generator automatically which works
                   # in a similar way. This is much more efficient.

Calling functions

dates_list = return_dates()
print(dates_list)
for i in dates_list:
    print(i)

dates_generator = yield_dates()
print(dates_generator)
for i in dates_generator:
    print(i)

Both functions do the same thing, but yield uses three lines instead of five and has one less variable to worry about.

This is the result from the code:

As you can see both functions do the same thing. The only difference is return_dates() gives a list and yield_dates() gives a generator.

A real life example would be something like reading a file line by line or if you just want to make a generator.

All great answers, however a bit difficult for newbies.

I assume you have learned the return statement.

As an analogy, return and yield are twins. return means 'return and stop' whereas 'yield` means 'return, but continue'

1. Try to get a num_list with return.

def num_list(n):
    for i in range(n):
        return i

Run it:

In [5]: num_list(3)
Out[5]: 0

See, you get only a single number rather than a list of them. return never allows you prevail happily, just implements once and quit.

2. There comes yield

Replace return with yield:

In [10]: def num_list(n):
    ...:     for i in range(n):
    ...:         yield i
    ...:

In [11]: num_list(3)
Out[11]: <generator object num_list at 0x10327c990>

In [12]: list(num_list(3))
Out[12]: [0, 1, 2]

Now, you win to get all the numbers.

Comparing to return which runs once and stops, yield runs times you planed. You can interpret return as return one of them, and yield as return all of them. This is called iterable.

3. One more step we can rewrite yield statement with return

In [15]: def num_list(n):
    ...:     result = []
    ...:     for i in range(n):
    ...:         result.append(i)
    ...:     return result

In [16]: num_list(3)
Out[16]: [0, 1, 2]

It's the core about yield.

The difference between a list return outputs and the object yield output is:

You will always get [0, 1, 2] from a list object but only could retrieve them from 'the object yield output' once. So, it has a new name generator object as displayed in Out[11]: <generator object num_list at 0x10327c990>.

In conclusion, as a metaphor to grok it:

• return and yield are twins
• list and generator are twins

Like every answer suggests, yield is used for creating a sequence generator. It's used for generating some sequence dynamically. For example, while reading a file line by line on a network, you can use the yield function as follows:

def getNextLines():
   while con.isOpen():
       yield con.read()

You can use it in your code as follows:

for line in getNextLines():
    doSomeThing(line)

Execution Control Transfer gotcha

The execution control will be transferred from getNextLines() to the for loop when yield is executed. Thus, every time getNextLines() is invoked, execution begins from the point where it was paused last time.

Thus in short, a function with the following code

def simpleYield():
    yield "first time"
    yield "second time"
    yield "third time"
    yield "Now some useful value {}".format(12)

for i in simpleYield():
    print i

will print

"first time"
"second time"
"third time"
"Now some useful value 12"

Yield

>>> def create_generator():
...    my_list = range(3)
...    for i in my_list:
...        yield i*i
...
>>> my_generator = create_generator() # create a generator
>>> print(my_generator) # my_generator is an object!
<generator object create_generator at 0xb7555c34>
>>> for i in my_generator:
...     print(i)
0
1
4

In short, you can see that the loop does not stop and continues to function even after the object or variable is sent (unlike return where the loop stops after execution).

Think of it this way:

An iterator is just a fancy sounding term for an object that has a next() method. So a yield-ed function ends up being something like this:

Original version:

def some_function():
    for i in xrange(4):
        yield i

for i in some_function():
    print i

This is basically what the Python interpreter does with the above code:

class it:
    def __init__(self):
        # Start at -1 so that we get 0 when we add 1 below.
        self.count = -1

    # The __iter__ method will be called once by the 'for' loop.
    # The rest of the magic happens on the object returned by this method.
    # In this case it is the object itself.
    def __iter__(self):
        return self

    # The next method will be called repeatedly by the 'for' loop
    # until it raises StopIteration.
    def next(self):
        self.count += 1
        if self.count < 4:
            return self.count
        else:
            # A StopIteration exception is raised
            # to signal that the iterator is done.
            # This is caught implicitly by the 'for' loop.
            raise StopIteration

def some_func():
    return it()

for i in some_func():
    print i

For more insight as to what's happening behind the scenes, the for loop can be rewritten to this:

iterator = some_func()
try:
    while 1:
        print iterator.next()
except StopIteration:
    pass

Does that make more sense or just confuse you more? :)

I should note that this is an oversimplification for illustrative purposes. :)

(My below answer only speaks from the perspective of using Python generator, not the underlying implementation of generator mechanism, which involves some tricks of stack and heap manipulation.)

When yield is used instead of a return in a python function, that function is turned into something special called generator function. That function will return an object of generator type. The yield keyword is a flag to notify the python compiler to treat such function specially. Normal functions will terminate once some value is returned from it. But with the help of the compiler, the generator function can be thought of as resumable. That is, the execution context will be restored and the execution will continue from last run. Until you explicitly call return, which will raise a StopIteration exception (which is also part of the iterator protocol), or reach the end of the function. I found a lot of references about generator but this one from the functional programming perspective is the most digestable.

(Now I want to talk about the rationale behind generator, and the iterator based on my own understanding. I hope this can help you grasp the essential motivation of iterator and generator. Such concept shows up in other languages as well such as C#.)

As I understand, when we want to process a bunch of data, we usually first store the data somewhere and then process it one by one. But this naive approach is problematic. If the data volume is huge, it's expensive to store them as a whole beforehand. So instead of storing the data itself directly, why not store some kind of metadata indirectly, i.e. the logic how the data is computed.

There are 2 approaches to wrap such metadata.

1. The OO approach, we wrap the metadata as a class. This is the so-called iterator who implements the iterator protocol (i.e. the __next__(), and __iter__() methods). This is also the commonly seen iterator design pattern.
2. The functional approach, we wrap the metadata as a function. This is the so-called generator function. But under the hood, the returned generator object still IS-A iterator because it also implements the iterator protocol.

Either way, an iterator is created, i.e. some object that can give you the data you want. The OO approach may be a bit complex. Anyway, which one to use is up to you.

In summary, the yield statement transforms your function into a factory that produces a special object called a generator which wraps around the body of your original function. When the generator is iterated, it executes your function until it reaches the next yield then suspends execution and evaluates to the value passed to yield. It repeats this process on each iteration until the path of execution exits the function. For instance,

def simple_generator():
    yield 'one'
    yield 'two'
    yield 'three'

for i in simple_generator():
    print i

simply outputs

one
two
three

The power comes from using the generator with a loop that calculates a sequence, the generator executes the loop stopping each time to 'yield' the next result of the calculation, in this way it calculates a list on the fly, the benefit being the memory saved for especially large calculations

Say you wanted to create a your own range function that produces an iterable range of numbers, you could do it like so,

def myRangeNaive(i):
    n = 0
    range = []
    while n < i:
        range.append(n)
        n = n + 1
    return range

and use it like this;

for i in myRangeNaive(10):
    print i

But this is inefficient because

• You create an array that you only use once (this wastes memory)
• This code actually loops over that array twice! :(

Luckily Guido and his team were generous enough to develop generators so we could just do this;

def myRangeSmart(i):
    n = 0
    while n < i:
       yield n
       n = n + 1
    return

for i in myRangeSmart(10):
    print i

Now upon each iteration a function on the generator called next() executes the function until it either reaches a 'yield' statement in which it stops and 'yields' the value or reaches the end of the function. In this case on the first call, next() executes up to the yield statement and yield 'n', on the next call it will execute the increment statement, jump back to the 'while', evaluate it, and if true, it will stop and yield 'n' again, it will continue that way until the while condition returns false and the generator jumps to the end of the function.

What does the yield keyword do in Python?

Answer Outline/Summary

• A function with yield, when called, returns a Generator.
• Generators are iterators because they implement the iterator protocol, so you can iterate over them.
• A generator can also be sent information, making it conceptually a coroutine.
• In Python 3, you can delegate from one generator to another in both directions with yield from.
• (Appendix critiques a couple of answers, including the top one, and discusses the use of return in a generator.)

Generators:

yield is only legal inside of a function definition, and the inclusion of yield in a function definition makes it return a generator.

The idea for generators comes from other languages (see footnote 1) with varying implementations. In Python's Generators, the execution of the code is frozen at the point of the yield. When the generator is called (methods are discussed below) execution resumes and then freezes at the next yield.

yield provides an easy way of implementing the iterator protocol, defined by the following two methods: __iter__ and next (Python 2) or __next__ (Python 3). Both of those methods make an object an iterator that you could type-check with the Iterator Abstract Base Class from the collections module.

>>> def func():
...     yield 'I am'
...     yield 'a generator!'
... 
>>> type(func)                 # A function with yield is still a function
<type 'function'>
>>> gen = func()
>>> type(gen)                  # but it returns a generator
<type 'generator'>
>>> hasattr(gen, '__iter__')   # that's an iterable
True
>>> hasattr(gen, 'next')       # and with .next (.__next__ in Python 3)
True                           # implements the iterator protocol.

The generator type is a sub-type of iterator:

>>> import collections, types
>>> issubclass(types.GeneratorType, collections.Iterator)
True

And if necessary, we can type-check like this:

>>> isinstance(gen, types.GeneratorType)
True
>>> isinstance(gen, collections.Iterator)
True

A feature of an Iterator is that once exhausted, you can't reuse or reset it:

>>> list(gen)
['I am', 'a generator!']
>>> list(gen)
[]

You'll have to make another if you want to use its functionality again (see footnote 2):

>>> list(func())
['I am', 'a generator!']

One can yield data programmatically, for example:

def func(an_iterable):
    for item in an_iterable:
        yield item

The above simple generator is also equivalent to the below - as of Python 3.3 (and not available in Python 2), you can use yield from:

def func(an_iterable):
    yield from an_iterable

However, yield from also allows for delegation to subgenerators, which will be explained in the following section on cooperative delegation with sub-coroutines.

Coroutines:

yield forms an expression that allows data to be sent into the generator (see footnote 3)

Here is an example, take note of the received variable, which will point to the data that is sent to the generator:

def bank_account(deposited, interest_rate):
    while True:
        calculated_interest = interest_rate * deposited 
        received = yield calculated_interest
        if received:
            deposited += received


>>> my_account = bank_account(1000, .05)

First, we must queue up the generator with the builtin function, next. It will call the appropriate next or __next__ method, depending on the version of Python you are using:

>>> first_year_interest = next(my_account)
>>> first_year_interest
50.0

And now we can send data into the generator. (Sending None is the same as calling next.) :

>>> next_year_interest = my_account.send(first_year_interest + 1000)
>>> next_year_interest
102.5

Cooperative Delegation to Sub-Coroutine with yield from

Now, recall that yield from is available in Python 3. This allows us to delegate coroutines to a subcoroutine:

def money_manager(expected_rate):
    # must receive deposited value from .send():
    under_management = yield                   # yield None to start.
    while True:
        try:
            additional_investment = yield expected_rate * under_management 
            if additional_investment:
                under_management += additional_investment
        except GeneratorExit:
            '''TODO: write function to send unclaimed funds to state'''
            raise
        finally:
            '''TODO: write function to mail tax info to client'''
        

def investment_account(deposited, manager):
    '''very simple model of an investment account that delegates to a manager'''
    # must queue up manager:
    next(manager)      # <- same as manager.send(None)
    # This is where we send the initial deposit to the manager:
    manager.send(deposited)
    try:
        yield from manager
    except GeneratorExit:
        return manager.close()  # delegate?

And now we can delegate functionality to a sub-generator and it can be used by a generator just as above:

my_manager = money_manager(.06)
my_account = investment_account(1000, my_manager)
first_year_return = next(my_account) # -> 60.0

Now simulate adding another 1,000 to the account plus the return on the account (60.0):

next_year_return = my_account.send(first_year_return + 1000)
next_year_return # 123.6

You can read more about the precise semantics of yield from in PEP 380.

Other Methods: close and throw

The close method raises GeneratorExit at the point the function execution was frozen. This will also be called by __del__ so you can put any cleanup code where you handle the GeneratorExit:

my_account.close()

You can also throw an exception which can be handled in the generator or propagated back to the user:

import sys
try:
    raise ValueError
except:
    my_manager.throw(*sys.exc_info())

Raises:

Traceback (most recent call last):
  File "<stdin>", line 4, in <module>
  File "<stdin>", line 6, in money_manager
  File "<stdin>", line 2, in <module>
ValueError

Conclusion

I believe I have covered all aspects of the following question:

What does the yield keyword do in Python?

It turns out that yield does a lot. I'm sure I could add even more thorough examples to this. If you want more or have some constructive criticism, let me know by commenting below.

Appendix:

Critique of the Top/Accepted Answer**

• It is confused on what makes an iterable, just using a list as an example. See my references above, but in summary: an iterable has an __iter__ method returning an iterator. An iterator provides a .next (Python 2 or .__next__ (Python 3) method, which is implicitly called by for loops until it raises StopIteration, and once it does, it will continue to do so.
• It then uses a generator expression to describe what a generator is. Since a generator is simply a convenient way to create an iterator, it only confuses the matter, and we still have not yet gotten to the yield part.
• In Controlling a generator exhaustion he calls the .next method, when instead he should use the builtin function, next. It would be an appropriate layer of indirection, because his code does not work in Python 3.
• Itertools? This was not relevant to what yield does at all.
• No discussion of the methods that yield provides along with the new functionality yield from in Python 3. The top/accepted answer is a very incomplete answer.

Critique of answer suggesting yield in a generator expression or comprehension.

The grammar currently allows any expression in a list comprehension.

expr_stmt: testlist_star_expr (annassign | augassign (yield_expr|testlist) |
                     ('=' (yield_expr|testlist_star_expr))*)
...
yield_expr: 'yield' [yield_arg]
yield_arg: 'from' test | testlist

Since yield is an expression, it has been touted by some as interesting to use it in comprehensions or generator expression - in spite of citing no particularly good use-case.

The CPython core developers are discussing deprecating its allowance. Here's a relevant post from the mailing list:

On 30 January 2017 at 19:05, Brett Cannon wrote:

On Sun, 29 Jan 2017 at 16:39 Craig Rodrigues wrote:

I'm OK with either approach. Leaving things the way they are in Python 3 is no good, IMHO.

My vote is it be a SyntaxError since you're not getting what you expect from the syntax.

I'd agree that's a sensible place for us to end up, as any code relying on the current behaviour is really too clever to be maintainable.

In terms of getting there, we'll likely want:

• SyntaxWarning or DeprecationWarning in 3.7
• Py3k warning in 2.7.x
• SyntaxError in 3.8

Cheers, Nick.

-- Nick Coghlan | ncoghlan at gmail.com | Brisbane, Australia

Further, there is an outstanding issue (10544) which seems to be pointing in the direction of this never being a good idea (PyPy, a Python implementation written in Python, is already raising syntax warnings.)

Bottom line, until the developers of CPython tell us otherwise: Don't put yield in a generator expression or comprehension.

The return statement in a generator

In Python 2:

In a generator function, the return statement is not allowed to include an expression_list. In that context, a bare return indicates that the generator is done and will cause StopIteration to be raised.

An expression_list is basically any number of expressions separated by commas - essentially, in Python 2, you can stop the generator with return, but you can't return a value.

In Python 3:

In a generator function, the return statement indicates that the generator is done and will cause StopIteration to be raised. The returned value (if any) is used as an argument to construct StopIteration and becomes the StopIteration.value attribute.

Footnotes

1. _{The languages CLU, Sather, and Icon were referenced in the proposal to introduce the concept of generators to Python. The general idea is that a function can maintain internal state and yield intermediate data points on demand by the user. This promised to be superior in performance to other approaches, including Python threading, which isn't even available on some systems.}

2. _{This means, for example, that range objects aren't Iterators, even though they are iterable, because they can be reused. Like lists, their __iter__ methods return iterator objects.}

yield was originally introduced as a statement, meaning that it could only appear at the beginning of a line in a code block. Now yield creates a yield expression. https://docs.python.org/2/reference/simple_stmts.html#grammar-token-yield_stmt This change was proposed to allow a user to send data into the generator just as one might receive it. To send data, one must be able to assign it to something, and for that, a statement just won't work.

An easy example to understand what it is: yield

def f123():
    for _ in range(4):
        yield 1
        yield 2


for i in f123():
    print (i)

The output is:

1 2 1 2 1 2 1 2

The yield keyword simply collects returning results. Think of yield like return +=

Yield gives you a generator.

def get_odd_numbers(i):
    return range(1, i, 2)
def yield_odd_numbers(i):
    for x in range(1, i, 2):
       yield x
foo = get_odd_numbers(10)
bar = yield_odd_numbers(10)
foo
[1, 3, 5, 7, 9]
bar
<generator object yield_odd_numbers at 0x1029c6f50>
bar.next()
1
bar.next()
3
bar.next()
5

As you can see, in the first case foo holds the entire list in memory at once. It's not a big deal for a list with 5 elements, but what if you want a list of 5 million? Not only is this a huge memory eater, it also costs a lot of time to build at the time that the function is called.

In the second case, bar just gives you a generator. A generator is an iterable--which means you can use it in a for loop, etc, but each value can only be accessed once. All the values are also not stored in memory at the same time; the generator object "remembers" where it was in the looping the last time you called it--this way, if you're using an iterable to (say) count to 50 billion, you don't have to count to 50 billion all at once and store the 50 billion numbers to count through.

Again, this is a pretty contrived example, you probably would use itertools if you really wanted to count to 50 billion. :)

This is the most simple use case of generators. As you said, it can be used to write efficient permutations, using yield to push things up through the call stack instead of using some sort of stack variable. Generators can also be used for specialized tree traversal, and all manner of other things.

Here is a simple example:

def isPrimeNumber(n):
    print "isPrimeNumber({}) call".format(n)
    if n==1:
        return False
    for x in range(2,n):
        if n % x == 0:
            return False
    return True

def primes (n=1):
    while(True):
        print "loop step ---------------- {}".format(n)
        if isPrimeNumber(n): yield n
        n += 1

for n in primes():
    if n> 10:break
    print "wiriting result {}".format(n)

Output:

loop step ---------------- 1
isPrimeNumber(1) call
loop step ---------------- 2
isPrimeNumber(2) call
loop step ---------------- 3
isPrimeNumber(3) call
wiriting result 3
loop step ---------------- 4
isPrimeNumber(4) call
loop step ---------------- 5
isPrimeNumber(5) call
wiriting result 5
loop step ---------------- 6
isPrimeNumber(6) call
loop step ---------------- 7
isPrimeNumber(7) call
wiriting result 7
loop step ---------------- 8
isPrimeNumber(8) call
loop step ---------------- 9
isPrimeNumber(9) call
loop step ---------------- 10
isPrimeNumber(10) call
loop step ---------------- 11
isPrimeNumber(11) call

I am not a Python developer, but it looks to me yield holds the position of program flow and the next loop start from "yield" position. It seems like it is waiting at that position, and just before that, returning a value outside, and next time continues to work.

It seems to be an interesting and nice ability :D

Here is an example in plain language. I will provide a correspondence between high-level human concepts to low-level Python concepts.

I want to operate on a sequence of numbers, but I don't want to bother my self with the creation of that sequence, I want only to focus on the operation I want to do. So, I do the following:

• I call you and tell you that I want a sequence of numbers which is produced in a specific way, and I let you know what the algorithm is.
  This step corresponds to defining the generator function, i.e. the function containing a yield.
• Sometime later, I tell you, "OK, get ready to tell me the sequence of numbers".
  This step corresponds to calling the generator function which returns a generator object. Note that you don't tell me any numbers yet; you just grab your paper and pencil.
• I ask you, "tell me the next number", and you tell me the first number; after that, you wait for me to ask you for the next number. It's your job to remember where you were, what numbers you have already said, and what is the next number. I don't care about the details.
  This step corresponds to calling .next() on the generator object.
• … repeat previous step, until…
• eventually, you might come to an end. You don't tell me a number; you just shout, "hold your horses! I'm done! No more numbers!"
  This step corresponds to the generator object ending its job, and raising a StopIteration exception The generator function does not need to raise the exception. It's raised automatically when the function ends or issues a return.

This is what a generator does (a function that contains a yield); it starts executing, pauses whenever it does a yield, and when asked for a .next() value it continues from the point it was last. It fits perfectly by design with the iterator protocol of Python, which describes how to sequentially request values.

The most famous user of the iterator protocol is the for command in Python. So, whenever you do a:

for item in sequence:

it doesn't matter if sequence is a list, a string, a dictionary or a generator object like described above; the result is the same: you read items off a sequence one by one.

Note that defining a function which contains a yield keyword is not the only way to create a generator; it's just the easiest way to create one.

For more accurate information, read about iterator types, the yield statement and generators in the Python documentation.

yield is just like return - it returns whatever you tell it to (as a generator). The difference is that the next time you call the generator, execution starts from the last call to the yield statement. Unlike return, the stack frame is not cleaned up when a yield occurs, however control is transferred back to the caller, so its state will resume the next time the function is called.

In the case of your code, the function get_child_candidates is acting like an iterator so that when you extend your list, it adds one element at a time to the new list.

list.extend calls an iterator until it's exhausted. In the case of the code sample you posted, it would be much clearer to just return a tuple and append that to the list.

It's returning a generator. I'm not particularly familiar with Python, but I believe it's the same kind of thing as C#'s iterator blocks if you're familiar with those.

The key idea is that the compiler/interpreter/whatever does some trickery so that as far as the caller is concerned, they can keep calling next() and it will keep returning values - as if the generator method was paused. Now obviously you can't really "pause" a method, so the compiler builds a state machine for you to remember where you currently are and what the local variables etc look like. This is much easier than writing an iterator yourself.

There is one type of answer that I don't feel has been given yet, among the many great answers that describe how to use generators. Here is the programming language theory answer:

The yield statement in Python returns a generator. A generator in Python is a function that returns continuations (and specifically a type of coroutine, but continuations represent the more general mechanism to understand what is going on).

Continuations in programming languages theory are a much more fundamental kind of computation, but they are not often used, because they are extremely hard to reason about and also very difficult to implement. But the idea of what a continuation is, is straightforward: it is the state of a computation that has not yet finished. In this state, the current values of variables, the operations that have yet to be performed, and so on, are saved. Then at some point later in the program the continuation can be invoked, such that the program's variables are reset to that state and the operations that were saved are carried out.

Continuations, in this more general form, can be implemented in two ways. In the call/cc way, the program's stack is literally saved and then when the continuation is invoked, the stack is restored.

In continuation passing style (CPS), continuations are just normal functions (only in languages where functions are first class) which the programmer explicitly manages and passes around to subroutines. In this style, program state is represented by closures (and the variables that happen to be encoded in them) rather than variables that reside somewhere on the stack. Functions that manage control flow accept continuation as arguments (in some variations of CPS, functions may accept multiple continuations) and manipulate control flow by invoking them by simply calling them and returning afterwards. A very simple example of continuation passing style is as follows:

def save_file(filename):
  def write_file_continuation():
    write_stuff_to_file(filename)

  check_if_file_exists_and_user_wants_to_overwrite(write_file_continuation)

In this (very simplistic) example, the programmer saves the operation of actually writing the file into a continuation (which can potentially be a very complex operation with many details to write out), and then passes that continuation (i.e, as a first-class closure) to another operator which does some more processing, and then calls it if necessary. (I use this design pattern a lot in actual GUI programming, either because it saves me lines of code or, more importantly, to manage control flow after GUI events trigger.)

The rest of this post will, without loss of generality, conceptualize continuations as CPS, because it is a hell of a lot easier to understand and read.

Now let's talk about generators in Python. Generators are a specific subtype of continuation. Whereas continuations are able in general to save the state of a computation (i.e., the program's call stack), generators are only able to save the state of iteration over an iterator. Although, this definition is slightly misleading for certain use cases of generators. For instance:

def f():
  while True:
    yield 4

This is clearly a reasonable iterable whose behavior is well defined -- each time the generator iterates over it, it returns 4 (and does so forever). But it isn't probably the prototypical type of iterable that comes to mind when thinking of iterators (i.e., for x in collection: do_something(x)). This example illustrates the power of generators: if anything is an iterator, a generator can save the state of its iteration.

To reiterate: Continuations can save the state of a program's stack and generators can save the state of iteration. This means that continuations are more a lot powerful than generators, but also that generators are a lot, lot easier. They are easier for the language designer to implement, and they are easier for the programmer to use (if you have some time to burn, try to read and understand this page about continuations and call/cc).

But you could easily implement (and conceptualize) generators as a simple, specific case of continuation passing style:

Whenever yield is called, it tells the function to return a continuation. When the function is called again, it starts from wherever it left off. So, in pseudo-pseudocode (i.e., not pseudocode, but not code) the generator's next method is basically as follows:

class Generator():
  def __init__(self,iterable,generatorfun):
    self.next_continuation = lambda:generatorfun(iterable)

  def next(self):
    value, next_continuation = self.next_continuation()
    self.next_continuation = next_continuation
    return value

where the yield keyword is actually syntactic sugar for the real generator function, basically something like:

def generatorfun(iterable):
  if len(iterable) == 0:
    raise StopIteration
  else:
    return (iterable[0], lambda:generatorfun(iterable[1:]))

Remember that this is just pseudocode and the actual implementation of generators in Python is more complex. But as an exercise to understand what is going on, try to use continuation passing style to implement generator objects without use of the yield keyword.

Shortcut to understanding yield

When you see a function with yield statements, apply this easy trick to understand what will happen:

1. Insert a line result = [] at the start of the function.
2. Replace each yield expr with result.append(expr).
3. Insert a line return result at the bottom of the function.
4. Yay - no more yield statements! Read and figure out code.
5. Compare function to the original definition.

This trick may give you an idea of the logic behind the function, but what actually happens with yield is significantly different than what happens in the list based approach. In many cases, the yield approach will be a lot more memory efficient and faster too. In other cases, this trick will get you stuck in an infinite loop, even though the original function works just fine. Read on to learn more...

Don't confuse your Iterables, Iterators, and Generators

First, the iterator protocol - when you write

for x in mylist:
    ...loop body...

Python performs the following two steps:

1. Gets an iterator for mylist:

   Call iter(mylist) -> this returns an object with a next() method (or __next__() in Python 3).

   [This is the step most people forget to tell you about]

2. Uses the iterator to loop over items:

   Keep calling the next() method on the iterator returned from step 1. The return value from next() is assigned to x and the loop body is executed. If an exception StopIteration is raised from within next(), it means there are no more values in the iterator and the loop is exited.

The truth is Python performs the above two steps anytime it wants to loop over the contents of an object - so it could be a for loop, but it could also be code like otherlist.extend(mylist) (where otherlist is a Python list).

Here mylist is an iterable because it implements the iterator protocol. In a user-defined class, you can implement the __iter__() method to make instances of your class iterable. This method should return an iterator. An iterator is an object with a next() method. It is possible to implement both __iter__() and next() on the same class, and have __iter__() return self. This will work for simple cases, but not when you want two iterators looping over the same object at the same time.

So that's the iterator protocol, many objects implement this protocol:

1. Built-in lists, dictionaries, tuples, sets, files.
2. User-defined classes that implement __iter__().
3. Generators.

Note that a for loop doesn't know what kind of object it's dealing with - it just follows the iterator protocol, and is happy to get item after item as it calls next(). Built-in lists return their items one by one, dictionaries return the keys one by one, files return the lines one by one, etc. And generators return... well that's where yield comes in:

def f123():
    yield 1
    yield 2
    yield 3

for item in f123():
    print item

Instead of yield statements, if you had three return statements in f123() only the first would get executed, and the function would exit. But f123() is no ordinary function. When f123() is called, it does not return any of the values in the yield statements! It returns a generator object. Also, the function does not really exit - it goes into a suspended state. When the for loop tries to loop over the generator object, the function resumes from its suspended state at the very next line after the yield it previously returned from, executes the next line of code, in this case, a yield statement, and returns that as the next item. This happens until the function exits, at which point the generator raises StopIteration, and the loop exits.

So the generator object is sort of like an adapter - at one end it exhibits the iterator protocol, by exposing __iter__() and next() methods to keep the for loop happy. At the other end, however, it runs the function just enough to get the next value out of it, and puts it back in suspended mode.

Why Use Generators?

Usually, you can write code that doesn't use generators but implements the same logic. One option is to use the temporary list 'trick' I mentioned before. That will not work in all cases, for e.g. if you have infinite loops, or it may make inefficient use of memory when you have a really long list. The other approach is to implement a new iterable class SomethingIter that keeps the state in instance members and performs the next logical step in it's next() (or __next__() in Python 3) method. Depending on the logic, the code inside the next() method may end up looking very complex and be prone to bugs. Here generators provide a clean and easy solution.

Here is a mental image of what yield does.

I like to think of a thread as having a stack (even when it's not implemented that way).

When a normal function is called, it puts its local variables on the stack, does some computation, then clears the stack and returns. The values of its local variables are never seen again.

With a yield function, when its code begins to run (i.e. after the function is called, returning a generator object, whose next() method is then invoked), it similarly puts its local variables onto the stack and computes for a while. But then, when it hits the yield statement, before clearing its part of the stack and returning, it takes a snapshot of its local variables and stores them in the generator object. It also writes down the place where it's currently up to in its code (i.e. the particular yield statement).

So it's a kind of a frozen function that the generator is hanging onto.

When next() is called subsequently, it retrieves the function's belongings onto the stack and re-animates it. The function continues to compute from where it left off, oblivious to the fact that it had just spent an eternity in cold storage.

Compare the following examples:

def normalFunction():
    return
    if False:
        pass

def yielderFunction():
    return
    if False:
        yield 12

When we call the second function, it behaves very differently to the first. The yield statement might be unreachable, but if it's present anywhere, it changes the nature of what we're dealing with.

>>> yielderFunction()
<generator object yielderFunction at 0x07742D28>

Calling yielderFunction() doesn't run its code, but makes a generator out of the code. (Maybe it's a good idea to name such things with the yielder prefix for readability.)

>>> gen = yielderFunction()
>>> dir(gen)
['__class__',
 ...
 '__iter__',    #Returns gen itself, to make it work uniformly with containers
 ...            #when given to a for loop. (Containers return an iterator instead.)
 'close',
 'gi_code',
 'gi_frame',
 'gi_running',
 'next',        #The method that runs the function's body.
 'send',
 'throw']

The gi_code and gi_frame fields are where the frozen state is stored. Exploring them with dir(..), we can confirm that our mental model above is credible.

There is one type of answer that I don't feel has been given yet, among the many great answers that describe how to use generators. Here is the programming language theory answer:

The yield statement in Python returns a generator. A generator in Python is a function that returns continuations (and specifically a type of coroutine, but continuations represent the more general mechanism to understand what is going on).

Continuations in programming languages theory are a much more fundamental kind of computation, but they are not often used, because they are extremely hard to reason about and also very difficult to implement. But the idea of what a continuation is, is straightforward: it is the state of a computation that has not yet finished. In this state, the current values of variables, the operations that have yet to be performed, and so on, are saved. Then at some point later in the program the continuation can be invoked, such that the program's variables are reset to that state and the operations that were saved are carried out.

Continuations, in this more general form, can be implemented in two ways. In the call/cc way, the program's stack is literally saved and then when the continuation is invoked, the stack is restored.

In continuation passing style (CPS), continuations are just normal functions (only in languages where functions are first class) which the programmer explicitly manages and passes around to subroutines. In this style, program state is represented by closures (and the variables that happen to be encoded in them) rather than variables that reside somewhere on the stack. Functions that manage control flow accept continuation as arguments (in some variations of CPS, functions may accept multiple continuations) and manipulate control flow by invoking them by simply calling them and returning afterwards. A very simple example of continuation passing style is as follows:

def save_file(filename):
  def write_file_continuation():
    write_stuff_to_file(filename)

  check_if_file_exists_and_user_wants_to_overwrite(write_file_continuation)

In this (very simplistic) example, the programmer saves the operation of actually writing the file into a continuation (which can potentially be a very complex operation with many details to write out), and then passes that continuation (i.e, as a first-class closure) to another operator which does some more processing, and then calls it if necessary. (I use this design pattern a lot in actual GUI programming, either because it saves me lines of code or, more importantly, to manage control flow after GUI events trigger.)

The rest of this post will, without loss of generality, conceptualize continuations as CPS, because it is a hell of a lot easier to understand and read.

Now let's talk about generators in Python. Generators are a specific subtype of continuation. Whereas continuations are able in general to save the state of a computation (i.e., the program's call stack), generators are only able to save the state of iteration over an iterator. Although, this definition is slightly misleading for certain use cases of generators. For instance:

def f():
  while True:
    yield 4

This is clearly a reasonable iterable whose behavior is well defined -- each time the generator iterates over it, it returns 4 (and does so forever). But it isn't probably the prototypical type of iterable that comes to mind when thinking of iterators (i.e., for x in collection: do_something(x)). This example illustrates the power of generators: if anything is an iterator, a generator can save the state of its iteration.

To reiterate: Continuations can save the state of a program's stack and generators can save the state of iteration. This means that continuations are more a lot powerful than generators, but also that generators are a lot, lot easier. They are easier for the language designer to implement, and they are easier for the programmer to use (if you have some time to burn, try to read and understand this page about continuations and call/cc).

But you could easily implement (and conceptualize) generators as a simple, specific case of continuation passing style:

Whenever yield is called, it tells the function to return a continuation. When the function is called again, it starts from wherever it left off. So, in pseudo-pseudocode (i.e., not pseudocode, but not code) the generator's next method is basically as follows:

class Generator():
  def __init__(self,iterable,generatorfun):
    self.next_continuation = lambda:generatorfun(iterable)

  def next(self):
    value, next_continuation = self.next_continuation()
    self.next_continuation = next_continuation
    return value

where the yield keyword is actually syntactic sugar for the real generator function, basically something like:

def generatorfun(iterable):
  if len(iterable) == 0:
    raise StopIteration
  else:
    return (iterable[0], lambda:generatorfun(iterable[1:]))

Remember that this is just pseudocode and the actual implementation of generators in Python is more complex. But as an exercise to understand what is going on, try to use continuation passing style to implement generator objects without use of the yield keyword.

yield in python is in a way similar to the return statement, except for some differences. If multiple values have to be returned from a function, return statement will return all the values as a list and it has to be stored in the memory in the caller block. But what if we don't want to use extra memory? Instead, we want to get the value from the function when we need it. This is where yield comes in. Consider the following function :-

def fun():
   yield 1
   yield 2
   yield 3

And the caller is :-

def caller():
   print ('First value printing')
   print (fun())
   print ('Second value printing')
   print (fun())
   print ('Third value printing')
   print (fun())

The above code segment (caller function) when called, outputs :-

First value printing
1
Second value printing
2
Third value printing
3

As can be seen from above, yield returns a value to its caller, but when the function is called again, it doesn't start from the first statement, but from the statement right after the yield. In the above example, "First value printing" was printed and the function was called. 1 was returned and printed. Then "Second value printing" was printed and again fun() was called. Instead of printing 1 (the first statement), it returned 2, i.e., the statement just after yield 1. The same process is repeated further.

There's one extra thing to mention: a function that yields doesn't actually have to terminate. I've written code like this:

def fib():
    last, cur = 0, 1
    while True: 
        yield cur
        last, cur = cur, last + cur

Then I can use it in other code like this:

for f in fib():
    if some_condition: break
    coolfuncs(f);

It really helps simplify some problems, and makes some things easier to work with.

yield is just like return - it returns whatever you tell it to (as a generator). The difference is that the next time you call the generator, execution starts from the last call to the yield statement. Unlike return, the stack frame is not cleaned up when a yield occurs, however control is transferred back to the caller, so its state will resume the next time the function is called.

In the case of your code, the function get_child_candidates is acting like an iterator so that when you extend your list, it adds one element at a time to the new list.

list.extend calls an iterator until it's exhausted. In the case of the code sample you posted, it would be much clearer to just return a tuple and append that to the list.

yield yields something. It's like somebody asks you to make 5 cupcakes. If you are done with at least one cupcake, you can give it to them to eat while you make other cakes.

In [4]: def make_cake(numbers):
   ...:     for i in range(numbers):
   ...:         yield 'Cake {}'.format(i)
   ...:

In [5]: factory = make_cake(5)

Here factory is called a generator, which makes you cakes. If you call make_function, you get a generator instead of running that function. It is because when yield keyword is present in a function, it becomes a generator.

In [7]: next(factory)
Out[7]: 'Cake 0'

In [8]: next(factory)
Out[8]: 'Cake 1'

In [9]: next(factory)
Out[9]: 'Cake 2'

In [10]: next(factory)
Out[10]: 'Cake 3'

In [11]: next(factory)
Out[11]: 'Cake 4'

They consumed all the cakes, but they ask for one again.

In [12]: next(factory)
---------------------------------------------------------------------------
StopIteration                             Traceback (most recent call last)
<ipython-input-12-0f5c45da9774> in <module>
----> 1 next(factory)

StopIteration:

and they are being told to stop asking more. So once you consumed a generator you are done with it. You need to call make_cake again if you want more cakes. It is like placing another order for cupcakes.

In [13]: factory = make_cake(3)

In [14]: for cake in factory:
    ...:     print(cake)
    ...:
Cake 0
Cake 1
Cake 2

You can also use for loop with a generator like the one above.

One more example: Lets say you want a random password whenever you ask for it.

In [22]: import random

In [23]: import string

In [24]: def random_password_generator():
    ...:     while True:
    ...:         yield ''.join([random.choice(string.ascii_letters) for _ in range(8)])
    ...:

In [25]: rpg = random_password_generator()

In [26]: for i in range(3):
    ...:     print(next(rpg))
    ...:
FXpUBhhH
DdUDHoHn
dvtebEqG

In [27]: next(rpg)
Out[27]: 'mJbYRMNo'

Here rpg is a generator, which can generate an infinite number of random passwords. So we can also say that generators are useful when we don't know the length of the sequence, unlike list which has a finite number of elements.

There's one extra thing to mention: a function that yields doesn't actually have to terminate. I've written code like this:

def fib():
    last, cur = 0, 1
    while True: 
        yield cur
        last, cur = cur, last + cur

Then I can use it in other code like this:

for f in fib():
    if some_condition: break
    coolfuncs(f);

It really helps simplify some problems, and makes some things easier to work with.

Here are some Python examples of how to actually implement generators as if Python did not provide syntactic sugar for them:

As a Python generator:

from itertools import islice

def fib_gen():
    a, b = 1, 1
    while True:
        yield a
        a, b = b, a + b

assert [1, 1, 2, 3, 5] == list(islice(fib_gen(), 5))

Using lexical closures instead of generators

def ftake(fnext, last):
    return [fnext() for _ in xrange(last)]

def fib_gen2():
    #funky scope due to python2.x workaround
    #for python 3.x use nonlocal
    def _():
        _.a, _.b = _.b, _.a + _.b
        return _.a
    _.a, _.b = 0, 1
    return _

assert [1,1,2,3,5] == ftake(fib_gen2(), 5)

Using object closures instead of generators (because ClosuresAndObjectsAreEquivalent)

class fib_gen3:
    def __init__(self):
        self.a, self.b = 1, 1

    def __call__(self):
        r = self.a
        self.a, self.b = self.b, self.a + self.b
        return r

assert [1,1,2,3,5] == ftake(fib_gen3(), 5)

(My below answer only speaks from the perspective of using Python generator, not the underlying implementation of generator mechanism, which involves some tricks of stack and heap manipulation.)

When yield is used instead of a return in a python function, that function is turned into something special called generator function. That function will return an object of generator type. The yield keyword is a flag to notify the python compiler to treat such function specially. Normal functions will terminate once some value is returned from it. But with the help of the compiler, the generator function can be thought of as resumable. That is, the execution context will be restored and the execution will continue from last run. Until you explicitly call return, which will raise a StopIteration exception (which is also part of the iterator protocol), or reach the end of the function. I found a lot of references about generator but this one from the functional programming perspective is the most digestable.

(Now I want to talk about the rationale behind generator, and the iterator based on my own understanding. I hope this can help you grasp the essential motivation of iterator and generator. Such concept shows up in other languages as well such as C#.)

As I understand, when we want to process a bunch of data, we usually first store the data somewhere and then process it one by one. But this naive approach is problematic. If the data volume is huge, it's expensive to store them as a whole beforehand. So instead of storing the data itself directly, why not store some kind of metadata indirectly, i.e. the logic how the data is computed.

There are 2 approaches to wrap such metadata.

1. The OO approach, we wrap the metadata as a class. This is the so-called iterator who implements the iterator protocol (i.e. the __next__(), and __iter__() methods). This is also the commonly seen iterator design pattern.
2. The functional approach, we wrap the metadata as a function. This is the so-called generator function. But under the hood, the returned generator object still IS-A iterator because it also implements the iterator protocol.

Either way, an iterator is created, i.e. some object that can give you the data you want. The OO approach may be a bit complex. Anyway, which one to use is up to you.

Think of it this way:

An iterator is just a fancy sounding term for an object that has a next() method. So a yield-ed function ends up being something like this:

Original version:

def some_function():
    for i in xrange(4):
        yield i

for i in some_function():
    print i

This is basically what the Python interpreter does with the above code:

class it:
    def __init__(self):
        # Start at -1 so that we get 0 when we add 1 below.
        self.count = -1

    # The __iter__ method will be called once by the 'for' loop.
    # The rest of the magic happens on the object returned by this method.
    # In this case it is the object itself.
    def __iter__(self):
        return self

    # The next method will be called repeatedly by the 'for' loop
    # until it raises StopIteration.
    def next(self):
        self.count += 1
        if self.count < 4:
            return self.count
        else:
            # A StopIteration exception is raised
            # to signal that the iterator is done.
            # This is caught implicitly by the 'for' loop.
            raise StopIteration

def some_func():
    return it()

for i in some_func():
    print i

For more insight as to what's happening behind the scenes, the for loop can be rewritten to this:

iterator = some_func()
try:
    while 1:
        print iterator.next()
except StopIteration:
    pass

Does that make more sense or just confuse you more? :)

I should note that this is an oversimplification for illustrative purposes. :)

An easy example to understand what it is: yield

def f123():
    for _ in range(4):
        yield 1
        yield 2


for i in f123():
    print (i)

The output is:

1 2 1 2 1 2 1 2

The yield keyword is reduced to two simple facts:

1. If the compiler detects the yield keyword anywhere inside a function, that function no longer returns via the return statement. Instead, it immediately returns a lazy "pending list" object called a generator
2. A generator is iterable. What is an iterable? It's anything like a list or set or range or dict-view, with a built-in protocol for visiting each element in a certain order.

In a nutshell: a generator is a lazy, incrementally-pending list, and yield statements allow you to use function notation to program the list values the generator should incrementally spit out.

generator = myYieldingFunction(...)
x = list(generator)

   generator
       v
[x[0], ..., ???]

         generator
             v
[x[0], x[1], ..., ???]

               generator
                   v
[x[0], x[1], x[2], ..., ???]

                       StopIteration exception
[x[0], x[1], x[2]]     done

list==[x[0], x[1], x[2]]

Example

Let's define a function makeRange that's just like Python's range. Calling makeRange(n) RETURNS A GENERATOR:

def makeRange(n):
    # return 0,1,2,...,n-1
    i = 0
    while i < n:
        yield i
        i += 1

>>> makeRange(5)
<generator object makeRange at 0x19e4aa0>

To force the generator to immediately return its pending values, you can pass it into list() (just like you could any iterable):

>>> list(makeRange(5))
[0, 1, 2, 3, 4]

Comparing example to "just returning a list"

The above example can be thought of as merely creating a list which you append to and return:

# list-version                   #  # generator-version
def makeRange(n):                #  def makeRange(n):
    """return [0,1,2,...,n-1]""" #~     """return 0,1,2,...,n-1"""
    TO_RETURN = []               #>
    i = 0                        #      i = 0
    while i < n:                 #      while i < n:
        TO_RETURN += [i]         #~         yield i
        i += 1                   #          i += 1  ## indented
    return TO_RETURN             #>

>>> makeRange(5)
[0, 1, 2, 3, 4]

There is one major difference, though; see the last section.

How you might use generators

An iterable is the last part of a list comprehension, and all generators are iterable, so they're often used like so:

#                   _ITERABLE_
>>> [x+10 for x in makeRange(5)]
[10, 11, 12, 13, 14]

To get a better feel for generators, you can play around with the itertools module (be sure to use chain.from_iterable rather than chain when warranted). For example, you might even use generators to implement infinitely-long lazy lists like itertools.count(). You could implement your own def enumerate(iterable): zip(count(), iterable), or alternatively do so with the yield keyword in a while-loop.

Please note: generators can actually be used for many more things, such as implementing coroutines or non-deterministic programming or other elegant things. However, the "lazy lists" viewpoint I present here is the most common use you will find.

Behind the scenes

This is how the "Python iteration protocol" works. That is, what is going on when you do list(makeRange(5)). This is what I describe earlier as a "lazy, incremental list".

>>> x=iter(range(5))
>>> next(x)
0
>>> next(x)
1
>>> next(x)
2
>>> next(x)
3
>>> next(x)
4
>>> next(x)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
StopIteration

The built-in function next() just calls the objects .next() function, which is a part of the "iteration protocol" and is found on all iterators. You can manually use the next() function (and other parts of the iteration protocol) to implement fancy things, usually at the expense of readability, so try to avoid doing that...

Minutiae

Normally, most people would not care about the following distinctions and probably want to stop reading here.

In Python-speak, an iterable is any object which "understands the concept of a for-loop" like a list [1,2,3], and an iterator is a specific instance of the requested for-loop like [1,2,3].__iter__(). A generator is exactly the same as any iterator, except for the way it was written (with function syntax).

When you request an iterator from a list, it creates a new iterator. However, when you request an iterator from an iterator (which you would rarely do), it just gives you a copy of itself.

Thus, in the unlikely event that you are failing to do something like this...

> x = myRange(5)
> list(x)
[0, 1, 2, 3, 4]
> list(x)
[]

... then remember that a generator is an iterator; that is, it is one-time-use. If you want to reuse it, you should call myRange(...) again. If you need to use the result twice, convert the result to a list and store it in a variable x = list(myRange(5)). Those who absolutely need to clone a generator (for example, who are doing terrifyingly hackish metaprogramming) can use itertools.tee if absolutely necessary, since the copyable iterator Python PEP standards proposal has been deferred.

Shortcut to understanding yield

When you see a function with yield statements, apply this easy trick to understand what will happen:

1. Insert a line result = [] at the start of the function.
2. Replace each yield expr with result.append(expr).
3. Insert a line return result at the bottom of the function.
4. Yay - no more yield statements! Read and figure out code.
5. Compare function to the original definition.

This trick may give you an idea of the logic behind the function, but what actually happens with yield is significantly different than what happens in the list based approach. In many cases, the yield approach will be a lot more memory efficient and faster too. In other cases, this trick will get you stuck in an infinite loop, even though the original function works just fine. Read on to learn more...

Don't confuse your Iterables, Iterators, and Generators

First, the iterator protocol - when you write

for x in mylist:
    ...loop body...

Python performs the following two steps:

1. Gets an iterator for mylist:

   Call iter(mylist) -> this returns an object with a next() method (or __next__() in Python 3).

   [This is the step most people forget to tell you about]

2. Uses the iterator to loop over items:

   Keep calling the next() method on the iterator returned from step 1. The return value from next() is assigned to x and the loop body is executed. If an exception StopIteration is raised from within next(), it means there are no more values in the iterator and the loop is exited.

The truth is Python performs the above two steps anytime it wants to loop over the contents of an object - so it could be a for loop, but it could also be code like otherlist.extend(mylist) (where otherlist is a Python list).

Here mylist is an iterable because it implements the iterator protocol. In a user-defined class, you can implement the __iter__() method to make instances of your class iterable. This method should return an iterator. An iterator is an object with a next() method. It is possible to implement both __iter__() and next() on the same class, and have __iter__() return self. This will work for simple cases, but not when you want two iterators looping over the same object at the same time.

So that's the iterator protocol, many objects implement this protocol:

1. Built-in lists, dictionaries, tuples, sets, files.
2. User-defined classes that implement __iter__().
3. Generators.

Note that a for loop doesn't know what kind of object it's dealing with - it just follows the iterator protocol, and is happy to get item after item as it calls next(). Built-in lists return their items one by one, dictionaries return the keys one by one, files return the lines one by one, etc. And generators return... well that's where yield comes in:

def f123():
    yield 1
    yield 2
    yield 3

for item in f123():
    print item

Instead of yield statements, if you had three return statements in f123() only the first would get executed, and the function would exit. But f123() is no ordinary function. When f123() is called, it does not return any of the values in the yield statements! It returns a generator object. Also, the function does not really exit - it goes into a suspended state. When the for loop tries to loop over the generator object, the function resumes from its suspended state at the very next line after the yield it previously returned from, executes the next line of code, in this case, a yield statement, and returns that as the next item. This happens until the function exits, at which point the generator raises StopIteration, and the loop exits.

So the generator object is sort of like an adapter - at one end it exhibits the iterator protocol, by exposing __iter__() and next() methods to keep the for loop happy. At the other end, however, it runs the function just enough to get the next value out of it, and puts it back in suspended mode.

Why Use Generators?

Usually, you can write code that doesn't use generators but implements the same logic. One option is to use the temporary list 'trick' I mentioned before. That will not work in all cases, for e.g. if you have infinite loops, or it may make inefficient use of memory when you have a really long list. The other approach is to implement a new iterable class SomethingIter that keeps the state in instance members and performs the next logical step in it's next() (or __next__() in Python 3) method. Depending on the logic, the code inside the next() method may end up looking very complex and be prone to bugs. Here generators provide a clean and easy solution.

It's returning a generator. I'm not particularly familiar with Python, but I believe it's the same kind of thing as C#'s iterator blocks if you're familiar with those.

The key idea is that the compiler/interpreter/whatever does some trickery so that as far as the caller is concerned, they can keep calling next() and it will keep returning values - as if the generator method was paused. Now obviously you can't really "pause" a method, so the compiler builds a state machine for you to remember where you currently are and what the local variables etc look like. This is much easier than writing an iterator yourself.

There is another yield use and meaning (since Python 3.3):

yield from <expr>

From PEP 380 -- Syntax for Delegating to a Subgenerator:

A syntax is proposed for a generator to delegate part of its operations to another generator. This allows a section of code containing 'yield' to be factored out and placed in another generator. Additionally, the subgenerator is allowed to return with a value, and the value is made available to the delegating generator.

The new syntax also opens up some opportunities for optimisation when one generator re-yields values produced by another.

Moreover this will introduce (since Python 3.5):

async def new_coroutine(data):
   ...
   await blocking_action()

to avoid coroutines being confused with a regular generator (today yield is used in both).

The yield keyword simply collects returning results. Think of yield like return +=

Here is an example in plain language. I will provide a correspondence between high-level human concepts to low-level Python concepts.

I want to operate on a sequence of numbers, but I don't want to bother my self with the creation of that sequence, I want only to focus on the operation I want to do. So, I do the following:

• I call you and tell you that I want a sequence of numbers which is produced in a specific way, and I let you know what the algorithm is.
  This step corresponds to defining the generator function, i.e. the function containing a yield.
• Sometime later, I tell you, "OK, get ready to tell me the sequence of numbers".
  This step corresponds to calling the generator function which returns a generator object. Note that you don't tell me any numbers yet; you just grab your paper and pencil.
• I ask you, "tell me the next number", and you tell me the first number; after that, you wait for me to ask you for the next number. It's your job to remember where you were, what numbers you have already said, and what is the next number. I don't care about the details.
  This step corresponds to calling .next() on the generator object.
• … repeat previous step, until…
• eventually, you might come to an end. You don't tell me a number; you just shout, "hold your horses! I'm done! No more numbers!"
  This step corresponds to the generator object ending its job, and raising a StopIteration exception The generator function does not need to raise the exception. It's raised automatically when the function ends or issues a return.

This is what a generator does (a function that contains a yield); it starts executing, pauses whenever it does a yield, and when asked for a .next() value it continues from the point it was last. It fits perfectly by design with the iterator protocol of Python, which describes how to sequentially request values.

The most famous user of the iterator protocol is the for command in Python. So, whenever you do a:

for item in sequence:

it doesn't matter if sequence is a list, a string, a dictionary or a generator object like described above; the result is the same: you read items off a sequence one by one.

Note that defining a function which contains a yield keyword is not the only way to create a generator; it's just the easiest way to create one.

For more accurate information, read about iterator types, the yield statement and generators in the Python documentation.

Here is a mental image of what yield does.

I like to think of a thread as having a stack (even when it's not implemented that way).

When a normal function is called, it puts its local variables on the stack, does some computation, then clears the stack and returns. The values of its local variables are never seen again.

With a yield function, when its code begins to run (i.e. after the function is called, returning a generator object, whose next() method is then invoked), it similarly puts its local variables onto the stack and computes for a while. But then, when it hits the yield statement, before clearing its part of the stack and returning, it takes a snapshot of its local variables and stores them in the generator object. It also writes down the place where it's currently up to in its code (i.e. the particular yield statement).

So it's a kind of a frozen function that the generator is hanging onto.

When next() is called subsequently, it retrieves the function's belongings onto the stack and re-animates it. The function continues to compute from where it left off, oblivious to the fact that it had just spent an eternity in cold storage.

Compare the following examples:

def normalFunction():
    return
    if False:
        pass

def yielderFunction():
    return
    if False:
        yield 12

When we call the second function, it behaves very differently to the first. The yield statement might be unreachable, but if it's present anywhere, it changes the nature of what we're dealing with.

>>> yielderFunction()
<generator object yielderFunction at 0x07742D28>

Calling yielderFunction() doesn't run its code, but makes a generator out of the code. (Maybe it's a good idea to name such things with the yielder prefix for readability.)

>>> gen = yielderFunction()
>>> dir(gen)
['__class__',
 ...
 '__iter__',    #Returns gen itself, to make it work uniformly with containers
 ...            #when given to a for loop. (Containers return an iterator instead.)
 'close',
 'gi_code',
 'gi_frame',
 'gi_running',
 'next',        #The method that runs the function's body.
 'send',
 'throw']

The gi_code and gi_frame fields are where the frozen state is stored. Exploring them with dir(..), we can confirm that our mental model above is credible.

While a lot of answers show why you'd use a yield to create a generator, there are more uses for yield. It's quite easy to make a coroutine, which enables the passing of information between two blocks of code. I won't repeat any of the fine examples that have already been given about using yield to create a generator.

To help understand what a yield does in the following code, you can use your finger to trace the cycle through any code that has a yield. Every time your finger hits the yield, you have to wait for a next or a send to be entered. When a next is called, you trace through the code until you hit the yield… the code on the right of the yield is evaluated and returned to the caller… then you wait. When next is called again, you perform another loop through the code. However, you'll note that in a coroutine, yield can also be used with a send… which will send a value from the caller into the yielding function. If a send is given, then yield receives the value sent, and spits it out the left hand side… then the trace through the code progresses until you hit the yield again (returning the value at the end, as if next was called).

For example:

>>> def coroutine():
...     i = -1
...     while True:
...         i += 1
...         val = (yield i)
...         print("Received %s" % val)
...
>>> sequence = coroutine()
>>> sequence.next()
0
>>> sequence.next()
Received None
1
>>> sequence.send('hello')
Received hello
2
>>> sequence.close()

Here's a simple yield based approach, to compute the fibonacci series, explained:

def fib(limit=50):
    a, b = 0, 1
    for i in range(limit):
       yield b
       a, b = b, a+b

When you enter this into your REPL and then try and call it, you'll get a mystifying result:

>>> fib()
<generator object fib at 0x7fa38394e3b8>

This is because the presence of yield signaled to Python that you want to create a generator, that is, an object that generates values on demand.

So, how do you generate these values? This can either be done directly by using the built-in function next, or, indirectly by feeding it to a construct that consumes values.

Using the built-in next() function, you directly invoke .next/__next__, forcing the generator to produce a value:

>>> g = fib()
>>> next(g)
1
>>> next(g)
1
>>> next(g)
2
>>> next(g)
3
>>> next(g)
5

Indirectly, if you provide fib to a for loop, a list initializer, a tuple initializer, or anything else that expects an object that generates/produces values, you'll "consume" the generator until no more values can be produced by it (and it returns):

results = []
for i in fib(30):       # consumes fib
    results.append(i) 
# can also be accomplished with
results = list(fib(30)) # consumes fib

Similarly, with a tuple initializer:

>>> tuple(fib(5))       # consumes fib
(1, 1, 2, 3, 5)

A generator differs from a function in the sense that it is lazy. It accomplishes this by maintaining it's local state and allowing you to resume whenever you need to.

When you first invoke fib by calling it:

f = fib()

Python compiles the function, encounters the yield keyword and simply returns a generator object back at you. Not very helpful it seems.

When you then request it generates the first value, directly or indirectly, it executes all statements that it finds, until it encounters a yield, it then yields back the value you supplied to yield and pauses. For an example that better demonstrates this, let's use some print calls (replace with print "text" if on Python 2):

def yielder(value):
    """ This is an infinite generator. Only use next on it """ 
    while 1:
        print("I'm going to generate the value for you")
        print("Then I'll pause for a while")
        yield value
        print("Let's go through it again.")

Now, enter in the REPL:

>>> gen = yielder("Hello, yield!")

you have a generator object now waiting for a command for it to generate a value. Use next and see what get's printed:

>>> next(gen) # runs until it finds a yield
I'm going to generate the value for you
Then I'll pause for a while
'Hello, yield!'

The unquoted results are what's printed. The quoted result is what is returned from yield. Call next again now:

>>> next(gen) # continues from yield and runs again
Let's go through it again.
I'm going to generate the value for you
Then I'll pause for a while
'Hello, yield!'

The generator remembers it was paused at yield value and resumes from there. The next message is printed and the search for the yield statement to pause at it performed again (due to the while loop).

The function containing the yield keyword is a generator. When you call it, its formal parameters are bound to actual arguments, but its body isn't actually evaluated. Instead, a generator-iterator is returned. Each call to the generator-iterator's next() method performs another pass through the iterative algorithm. Each step's value is the value specified by the yield keyword. Think of yield as the generator-iterator version of return, indicating the boundary between each iteration of the algorithm. Each time you call next(), the generator code resumes from the statement following the yield.

Think of it this way:

An iterator is just a fancy sounding term for an object that has a next() method. So a yield-ed function ends up being something like this:

Original version:

def some_function():
    for i in xrange(4):
        yield i

for i in some_function():
    print i

This is basically what the Python interpreter does with the above code:

class it:
    def __init__(self):
        # Start at -1 so that we get 0 when we add 1 below.
        self.count = -1

    # The __iter__ method will be called once by the 'for' loop.
    # The rest of the magic happens on the object returned by this method.
    # In this case it is the object itself.
    def __iter__(self):
        return self

    # The next method will be called repeatedly by the 'for' loop
    # until it raises StopIteration.
    def next(self):
        self.count += 1
        if self.count < 4:
            return self.count
        else:
            # A StopIteration exception is raised
            # to signal that the iterator is done.
            # This is caught implicitly by the 'for' loop.
            raise StopIteration

def some_func():
    return it()

for i in some_func():
    print i

For more insight as to what's happening behind the scenes, the for loop can be rewritten to this:

iterator = some_func()
try:
    while 1:
        print iterator.next()
except StopIteration:
    pass

Does that make more sense or just confuse you more? :)

I should note that this is an oversimplification for illustrative purposes. :)

Shortcut to understanding yield

When you see a function with yield statements, apply this easy trick to understand what will happen:

1. Insert a line result = [] at the start of the function.
2. Replace each yield expr with result.append(expr).
3. Insert a line return result at the bottom of the function.
4. Yay - no more yield statements! Read and figure out code.
5. Compare function to the original definition.

This trick may give you an idea of the logic behind the function, but what actually happens with yield is significantly different than what happens in the list based approach. In many cases, the yield approach will be a lot more memory efficient and faster too. In other cases, this trick will get you stuck in an infinite loop, even though the original function works just fine. Read on to learn more...

Don't confuse your Iterables, Iterators, and Generators

First, the iterator protocol - when you write

for x in mylist:
    ...loop body...

Python performs the following two steps:

1. Gets an iterator for mylist:

   Call iter(mylist) -> this returns an object with a next() method (or __next__() in Python 3).

   [This is the step most people forget to tell you about]

2. Uses the iterator to loop over items:

   Keep calling the next() method on the iterator returned from step 1. The return value from next() is assigned to x and the loop body is executed. If an exception StopIteration is raised from within next(), it means there are no more values in the iterator and the loop is exited.

The truth is Python performs the above two steps anytime it wants to loop over the contents of an object - so it could be a for loop, but it could also be code like otherlist.extend(mylist) (where otherlist is a Python list).

Here mylist is an iterable because it implements the iterator protocol. In a user-defined class, you can implement the __iter__() method to make instances of your class iterable. This method should return an iterator. An iterator is an object with a next() method. It is possible to implement both __iter__() and next() on the same class, and have __iter__() return self. This will work for simple cases, but not when you want two iterators looping over the same object at the same time.

So that's the iterator protocol, many objects implement this protocol:

1. Built-in lists, dictionaries, tuples, sets, files.
2. User-defined classes that implement __iter__().
3. Generators.

Note that a for loop doesn't know what kind of object it's dealing with - it just follows the iterator protocol, and is happy to get item after item as it calls next(). Built-in lists return their items one by one, dictionaries return the keys one by one, files return the lines one by one, etc. And generators return... well that's where yield comes in:

def f123():
    yield 1
    yield 2
    yield 3

for item in f123():
    print item

Instead of yield statements, if you had three return statements in f123() only the first would get executed, and the function would exit. But f123() is no ordinary function. When f123() is called, it does not return any of the values in the yield statements! It returns a generator object. Also, the function does not really exit - it goes into a suspended state. When the for loop tries to loop over the generator object, the function resumes from its suspended state at the very next line after the yield it previously returned from, executes the next line of code, in this case, a yield statement, and returns that as the next item. This happens until the function exits, at which point the generator raises StopIteration, and the loop exits.

So the generator object is sort of like an adapter - at one end it exhibits the iterator protocol, by exposing __iter__() and next() methods to keep the for loop happy. At the other end, however, it runs the function just enough to get the next value out of it, and puts it back in suspended mode.

Why Use Generators?

Usually, you can write code that doesn't use generators but implements the same logic. One option is to use the temporary list 'trick' I mentioned before. That will not work in all cases, for e.g. if you have infinite loops, or it may make inefficient use of memory when you have a really long list. The other approach is to implement a new iterable class SomethingIter that keeps the state in instance members and performs the next logical step in it's next() (or __next__() in Python 3) method. Depending on the logic, the code inside the next() method may end up looking very complex and be prone to bugs. Here generators provide a clean and easy solution.

A simple generator function

def my_gen():
    n = 1
    print('This is printed first')
    # Generator function contains yield statements
    yield n

    n += 1
    print('This is printed second')
    yield n

    n += 1
    print('This is printed at last')
    yield n

yield statement pauses the function saving all its states and later continues from there on successive calls.

https://www.programiz.com/python-programming/generator

From a programming viewpoint, the iterators are implemented as thunks.

To implement iterators, generators, and thread pools for concurrent execution, etc. as thunks, one uses messages sent to a closure object, which has a dispatcher, and the dispatcher answers to "messages".

"next" is a message sent to a closure, created by the "iter" call.

There are lots of ways to implement this computation. I used mutation, but it is possible to do this kind of computation without mutation, by returning the current value and the next yielder (making it referential transparent). Racket uses a sequence of transformations of the initial program in some intermediary languages, one of such rewriting making the yield operator to be transformed in some language with simpler operators.

Here is a demonstration of how yield could be rewritten, which uses the structure of R6RS, but the semantics is identical to Python's. It's the same model of computation, and only a change in syntax is required to rewrite it using yield of Python.

Welcome to Racket v6.5.0.3.

-> (define gen
     (lambda (l)
       (define yield
         (lambda ()
           (if (null? l)
               'END
               (let ((v (car l)))
                 (set! l (cdr l))
                 v))))
       (lambda(m)
         (case m
           ('yield (yield))
           ('init  (lambda (data)
                     (set! l data)
                     'OK))))))
-> (define stream (gen '(1 2 3)))
-> (stream 'yield)
1
-> (stream 'yield)
2
-> (stream 'yield)
3
-> (stream 'yield)
'END
-> ((stream 'init) '(a b))
'OK
-> (stream 'yield)
'a
-> (stream 'yield)
'b
-> (stream 'yield)
'END
-> (stream 'yield)
'END
->

There is another yield use and meaning (since Python 3.3):

yield from <expr>

From PEP 380 -- Syntax for Delegating to a Subgenerator:

A syntax is proposed for a generator to delegate part of its operations to another generator. This allows a section of code containing 'yield' to be factored out and placed in another generator. Additionally, the subgenerator is allowed to return with a value, and the value is made available to the delegating generator.

The new syntax also opens up some opportunities for optimisation when one generator re-yields values produced by another.

Moreover this will introduce (since Python 3.5):

async def new_coroutine(data):
   ...
   await blocking_action()

to avoid coroutines being confused with a regular generator (today yield is used in both).

Can also send data back to the generator!

Indeed, as many answers here explain, using yield creates a generator.

You can use the yield keyword to send data back to a "live" generator.

Example:

Let's say we have a method which translates from english to some other language. And in the beginning of it, it does something which is heavy and should be done once. We want this method run forever (don't really know why.. :)), and receive words words to be translated.

def translator():
    # load all the words in English language and the translation to 'other lang'
    my_words_dict = {'hello': 'hello in other language', 'dog': 'dog in other language'}

    while True:
        word = (yield)
        yield my_words_dict.get(word, 'Unknown word...')

Running:

my_words_translator = translator()

next(my_words_translator)
print(my_words_translator.send('dog'))

next(my_words_translator)
print(my_words_translator.send('cat'))

will print:

dog in other language
Unknown word...

To summarise:

use send method inside a generator to send data back to the generator. To allow that, a (yield) is used.

All of the answers here are great; but only one of them (the most voted one) relates to how your code works. Others are relating to generators in general, and how they work.

So I won't repeat what generators are or what yields do; I think these are covered by great existing answers. However, after spending few hours trying to understand a similar code to yours, I'll break it down how it works.

Your code traverse a binary tree structure. Let's take this tree for example:

    5
   / \
  3   6
 / \   \
1   4   8

And another simpler implementation of a binary-search tree traversal:

class Node(object):
..
def __iter__(self):
    if self.has_left_child():
        for child in self.left:
            yield child

    yield self.val

    if self.has_right_child():
        for child in self.right:
            yield child

The execution code is on the Tree object, which implements __iter__ as this:

def __iter__(self):

    class EmptyIter():
        def next(self):
            raise StopIteration

    if self.root:
        return self.root.__iter__()
    return EmptyIter()

The while candidates statement can be replaced with for element in tree; Python translate this to

it = iter(TreeObj)  # returns iter(self.root) which calls self.root.__iter__()
for element in it: 
    .. process element .. 

Because Node.__iter__ function is a generator, the code inside it is executed per iteration. So the execution would look like this:

1. root element is first; check if it has left childs and for iterate them (let's call it it1 because its the first iterator object)
2. it has a child so the for is executed. The for child in self.left creates a new iterator from self.left, which is a Node object itself (it2)
3. Same logic as 2, and a new iterator is created (it3)
4. Now we reached the left end of the tree. it3 has no left childs so it continues and yield self.value
5. On the next call to next(it3) it raises StopIteration and exists since it has no right childs (it reaches to the end of the function without yield anything)
6. it1 and it2 are still active - they are not exhausted and calling next(it2) would yield values, not raise StopIteration
7. Now we are back to it2 context, and call next(it2) which continues where it stopped: right after the yield child statement. Since it has no more left childs it continues and yields it's self.val.

The catch here is that every iteration creates sub-iterators to traverse the tree, and holds the state of the current iterator. Once it reaches the end it traverse back the stack, and values are returned in the correct order (smallest yields value first).

Your code example did something similar in a different technique: it populated a one-element list for every child, then on the next iteration it pops it and run the function code on the current object (hence the self).

I hope this contributed a little to this legendary topic. I spent several good hours drawing this process to understand it.

TL;DR

Instead of this:

def square_list(n):
    the_list = []                         # Replace
    for x in range(n):
        y = x * x
        the_list.append(y)                # these
    return the_list                       # lines

do this:

def square_yield(n):
    for x in range(n):
        y = x * x
        yield y                           # with this one.

Whenever you find yourself building a list from scratch, yield each piece instead.

This was my first "aha" moment with yield.

yield is a sugary way to say

build a series of stuff

Same behavior:

>>> for square in square_list(4):
...     print(square)
...
0
1
4
9
>>> for square in square_yield(4):
...     print(square)
...
0
1
4
9

Different behavior:

Yield is single-pass: you can only iterate through once. When a function has a yield in it we call it a generator function. And an iterator is what it returns. Those terms are revealing. We lose the convenience of a container, but gain the power of a series that's computed as needed, and arbitrarily long.

Yield is lazy, it puts off computation. A function with a yield in it doesn't actually execute at all when you call it. It returns an iterator object that remembers where it left off. Each time you call next() on the iterator (this happens in a for-loop) execution inches forward to the next yield. return raises StopIteration and ends the series (this is the natural end of a for-loop).

Yield is versatile. Data doesn't have to be stored all together, it can be made available one at a time. It can be infinite.

>>> def squares_all_of_them():
...     x = 0
...     while True:
...         yield x * x
...         x += 1
...
>>> squares = squares_all_of_them()
>>> for _ in range(4):
...     print(next(squares))
...
0
1
4
9

If you need multiple passes and the series isn't too long, just call list() on it:

>>> list(square_yield(4))
[0, 1, 4, 9]

Brilliant choice of the word yield because both meanings apply:

yield — produce or provide (as in agriculture)

...provide the next data in the series.

yield — give way or relinquish (as in political power)

...relinquish CPU execution until the iterator advances.

yield is like a return element for a function. The difference is, that the yield element turns a function into a generator. A generator behaves just like a function until something is 'yielded'. The generator stops until it is next called, and continues from exactly the same point as it started. You can get a sequence of all the 'yielded' values in one, by calling list(generator()).

A simple generator function

def my_gen():
    n = 1
    print('This is printed first')
    # Generator function contains yield statements
    yield n

    n += 1
    print('This is printed second')
    yield n

    n += 1
    print('This is printed at last')
    yield n

yield statement pauses the function saving all its states and later continues from there on successive calls.

https://www.programiz.com/python-programming/generator

For those who prefer a minimal working example, meditate on this interactive Python session:

>>> def f():
...   yield 1
...   yield 2
...   yield 3
... 
>>> g = f()
>>> for i in g:
...   print(i)
... 
1
2
3
>>> for i in g:
...   print(i)
... 
>>> # Note that this time nothing was printed

From a programming viewpoint, the iterators are implemented as thunks.

To implement iterators, generators, and thread pools for concurrent execution, etc. as thunks, one uses messages sent to a closure object, which has a dispatcher, and the dispatcher answers to "messages".

"next" is a message sent to a closure, created by the "iter" call.

There are lots of ways to implement this computation. I used mutation, but it is possible to do this kind of computation without mutation, by returning the current value and the next yielder (making it referential transparent). Racket uses a sequence of transformations of the initial program in some intermediary languages, one of such rewriting making the yield operator to be transformed in some language with simpler operators.

Here is a demonstration of how yield could be rewritten, which uses the structure of R6RS, but the semantics is identical to Python's. It's the same model of computation, and only a change in syntax is required to rewrite it using yield of Python.

Welcome to Racket v6.5.0.3.

-> (define gen
     (lambda (l)
       (define yield
         (lambda ()
           (if (null? l)
               'END
               (let ((v (car l)))
                 (set! l (cdr l))
                 v))))
       (lambda(m)
         (case m
           ('yield (yield))
           ('init  (lambda (data)
                     (set! l data)
                     'OK))))))
-> (define stream (gen '(1 2 3)))
-> (stream 'yield)
1
-> (stream 'yield)
2
-> (stream 'yield)
3
-> (stream 'yield)
'END
-> ((stream 'init) '(a b))
'OK
-> (stream 'yield)
'a
-> (stream 'yield)
'b
-> (stream 'yield)
'END
-> (stream 'yield)
'END
->

yield is similar to return. The difference is:

yield makes a function iterable (in the following example primes(n = 1) function becomes iterable).
What it essentially means is the next time the function is called, it will continue from where it left (which is after the line of yield expression).

def isprime(n):
    if n == 1:
        return False
    for x in range(2, n):
        if n % x == 0:
            return False
    else:
        return True

def primes(n = 1):
   while(True):
       if isprime(n): yield n
       n += 1 

for n in primes():
    if n > 100: break
    print(n)

In the above example if isprime(n) is true it will return the prime number. In the next iteration it will continue from the next line

n += 1  

TL;DR

Instead of this:

def square_list(n):
    the_list = []                         # Replace
    for x in range(n):
        y = x * x
        the_list.append(y)                # these
    return the_list                       # lines

do this:

def square_yield(n):
    for x in range(n):
        y = x * x
        yield y                           # with this one.

Whenever you find yourself building a list from scratch, yield each piece instead.

This was my first "aha" moment with yield.

yield is a sugary way to say

build a series of stuff

Same behavior:

>>> for square in square_list(4):
...     print(square)
...
0
1
4
9
>>> for square in square_yield(4):
...     print(square)
...
0
1
4
9

Different behavior:

Yield is single-pass: you can only iterate through once. When a function has a yield in it we call it a generator function. And an iterator is what it returns. Those terms are revealing. We lose the convenience of a container, but gain the power of a series that's computed as needed, and arbitrarily long.

Yield is lazy, it puts off computation. A function with a yield in it doesn't actually execute at all when you call it. It returns an iterator object that remembers where it left off. Each time you call next() on the iterator (this happens in a for-loop) execution inches forward to the next yield. return raises StopIteration and ends the series (this is the natural end of a for-loop).

Yield is versatile. Data doesn't have to be stored all together, it can be made available one at a time. It can be infinite.

>>> def squares_all_of_them():
...     x = 0
...     while True:
...         yield x * x
...         x += 1
...
>>> squares = squares_all_of_them()
>>> for _ in range(4):
...     print(next(squares))
...
0
1
4
9

If you need multiple passes and the series isn't too long, just call list() on it:

>>> list(square_yield(4))
[0, 1, 4, 9]

Brilliant choice of the word yield because both meanings apply:

yield — produce or provide (as in agriculture)

...provide the next data in the series.

yield — give way or relinquish (as in political power)

...relinquish CPU execution until the iterator advances.

Here is a simple example:

def isPrimeNumber(n):
    print "isPrimeNumber({}) call".format(n)
    if n==1:
        return False
    for x in range(2,n):
        if n % x == 0:
            return False
    return True

def primes (n=1):
    while(True):
        print "loop step ---------------- {}".format(n)
        if isPrimeNumber(n): yield n
        n += 1

for n in primes():
    if n> 10:break
    print "wiriting result {}".format(n)

Output:

loop step ---------------- 1
isPrimeNumber(1) call
loop step ---------------- 2
isPrimeNumber(2) call
loop step ---------------- 3
isPrimeNumber(3) call
wiriting result 3
loop step ---------------- 4
isPrimeNumber(4) call
loop step ---------------- 5
isPrimeNumber(5) call
wiriting result 5
loop step ---------------- 6
isPrimeNumber(6) call
loop step ---------------- 7
isPrimeNumber(7) call
wiriting result 7
loop step ---------------- 8
isPrimeNumber(8) call
loop step ---------------- 9
isPrimeNumber(9) call
loop step ---------------- 10
isPrimeNumber(10) call
loop step ---------------- 11
isPrimeNumber(11) call

I am not a Python developer, but it looks to me yield holds the position of program flow and the next loop start from "yield" position. It seems like it is waiting at that position, and just before that, returning a value outside, and next time continues to work.

It seems to be an interesting and nice ability :D

For those who prefer a minimal working example, meditate on this interactive Python session:

>>> def f():
...   yield 1
...   yield 2
...   yield 3
... 
>>> g = f()
>>> for i in g:
...   print(i)
... 
1
2
3
>>> for i in g:
...   print(i)
... 
>>> # Note that this time nothing was printed

yield is similar to return. The difference is:

yield makes a function iterable (in the following example primes(n = 1) function becomes iterable).
What it essentially means is the next time the function is called, it will continue from where it left (which is after the line of yield expression).

def isprime(n):
    if n == 1:
        return False
    for x in range(2, n):
        if n % x == 0:
            return False
    else:
        return True

def primes(n = 1):
   while(True):
       if isprime(n): yield n
       n += 1 

for n in primes():
    if n > 100: break
    print(n)

In the above example if isprime(n) is true it will return the prime number. In the next iteration it will continue from the next line

n += 1  

In Python generators (a special type of iterators) are used to generate series of values and yield keyword is just like the return keyword of generator functions.

The other fascinating thing yield keyword does is saving the state of a generator function.

So, we can set a number to a different value each time the generator yields.

Here's an instance:

def getPrimes(number):
    while True:
        if isPrime(number):
            number = yield number     # a miracle occurs here
        number += 1

def printSuccessivePrimes(iterations, base=10):
    primeGenerator = getPrimes(base)
    primeGenerator.send(None)
    for power in range(iterations):
        print(primeGenerator.send(base ** power))

An analogy could help to grasp the idea here:

Imagine that you have created an amazing machine that is capable of generating thousands and thousands of lightbulbs per day. The machine generates these lightbulbs in boxes with a unique serial number. You don't have enough space to store all these lightbulbs at the same time (i.e., you cannot keep up with the speed of the machine due to storage limitation), so you would like to adjust this machine to generate lightbulbs on demand.

Python generators don't differ much from this concept.

Imagine that you have a function x that generates unique serial numbers for the boxes. Obviously, you can have a very large number of such barcodes generated by the function. A wiser, and space efficient, option is to generate those serial numbers on-demand.

Machine's code:

def barcode_generator():
    serial_number = 10000  # Initial barcode
    while True:
        yield serial_number
        serial_number += 1


barcode = barcode_generator()
while True:
    number_of_lightbulbs_to_generate = int(input("How many lightbulbs to generate? "))
    barcodes = [next(barcode) for _ in range(number_of_lightbulbs_to_generate)]
    print(barcodes)

    # function_to_create_the_next_batch_of_lightbulbs(barcodes)

    produce_more = input("Produce more? [Y/n]: ")
    if produce_more == "n":
        break

As you can see we have a self-contained "function" to generate the next unique serial number each time. This function returns back a generator! As you can see we are not calling the function each time we need a new serial number, but we are using next() given the generator to obtain the next serial number.

Output:

How many lightbulbs to generate? 5
[10000, 10001, 10002, 10003, 10004]
Produce more? [Y/n]: y
How many lightbulbs to generate? 6
[10005, 10006, 10007, 10008, 10009, 10010]
Produce more? [Y/n]: y
How many lightbulbs to generate? 7
[10011, 10012, 10013, 10014, 10015, 10016, 10017]
Produce more? [Y/n]: n

There's one extra thing to mention: a function that yields doesn't actually have to terminate. I've written code like this:

def fib():
    last, cur = 0, 1
    while True: 
        yield cur
        last, cur = cur, last + cur

Then I can use it in other code like this:

for f in fib():
    if some_condition: break
    coolfuncs(f);

It really helps simplify some problems, and makes some things easier to work with.

The yield keyword is going to replace return in a function definition to create a generator.

def create_generator():
   for i in range(100):
   yield i
myGenerator = create_generator()
print(myGenerator)
# <generator object create_generator at 0x102dd2480>
for i in myGenerator:
   print(i) # prints 0-99

When the returned generator is first used—not in the assignment but the for loop—the function definition will execute until it reaches the yield statement. There, it will pause (see why it’s called yield) until used again. Then, it will pick up where it left off. Upon the final iteration of the generator, any code after the yield command will execute.

def create_generator():
   print("Beginning of generator")
   for i in range(4):
      yield i
   print("After yield")
print("Before assignment")

myGenerator = create_generator()

print("After assignment")
for i in myGenerator :
   print(i) # prints 0-3
"""
Before assignment
After assignment
Beginning of generator
0
1
2
After yield

The yield keyword modifies a function’s behavior to produce a generator that’s paused at each yield command during iteration. The function isn’t executed except upon iteration, which leads to improved resource management, and subsequently, a better overall performance. Use generators (and yielded functions) for creating large data sets meant for single-use iteration.

Think of it this way:

An iterator is just a fancy sounding term for an object that has a next() method. So a yield-ed function ends up being something like this:

Original version:

def some_function():
    for i in xrange(4):
        yield i

for i in some_function():
    print i

This is basically what the Python interpreter does with the above code:

class it:
    def __init__(self):
        # Start at -1 so that we get 0 when we add 1 below.
        self.count = -1

    # The __iter__ method will be called once by the 'for' loop.
    # The rest of the magic happens on the object returned by this method.
    # In this case it is the object itself.
    def __iter__(self):
        return self

    # The next method will be called repeatedly by the 'for' loop
    # until it raises StopIteration.
    def next(self):
        self.count += 1
        if self.count < 4:
            return self.count
        else:
            # A StopIteration exception is raised
            # to signal that the iterator is done.
            # This is caught implicitly by the 'for' loop.
            raise StopIteration

def some_func():
    return it()

for i in some_func():
    print i

For more insight as to what's happening behind the scenes, the for loop can be rewritten to this:

iterator = some_func()
try:
    while 1:
        print iterator.next()
except StopIteration:
    pass

Does that make more sense or just confuse you more? :)

I should note that this is an oversimplification for illustrative purposes. :)

The function containing the yield keyword is a generator. When you call it, its formal parameters are bound to actual arguments, but its body isn't actually evaluated. Instead, a generator-iterator is returned. Each call to the generator-iterator's next() method performs another pass through the iterative algorithm. Each step's value is the value specified by the yield keyword. Think of yield as the generator-iterator version of return, indicating the boundary between each iteration of the algorithm. Each time you call next(), the generator code resumes from the statement following the yield.

What does the yield keyword do in Python?

Answer Outline/Summary

• A function with yield, when called, returns a Generator.
• Generators are iterators because they implement the iterator protocol, so you can iterate over them.
• A generator can also be sent information, making it conceptually a coroutine.
• In Python 3, you can delegate from one generator to another in both directions with yield from.
• (Appendix critiques a couple of answers, including the top one, and discusses the use of return in a generator.)

Generators:

yield is only legal inside of a function definition, and the inclusion of yield in a function definition makes it return a generator.

The idea for generators comes from other languages (see footnote 1) with varying implementations. In Python's Generators, the execution of the code is frozen at the point of the yield. When the generator is called (methods are discussed below) execution resumes and then freezes at the next yield.

yield provides an easy way of implementing the iterator protocol, defined by the following two methods: __iter__ and next (Python 2) or __next__ (Python 3). Both of those methods make an object an iterator that you could type-check with the Iterator Abstract Base Class from the collections module.

>>> def func():
...     yield 'I am'
...     yield 'a generator!'
... 
>>> type(func)                 # A function with yield is still a function
<type 'function'>
>>> gen = func()
>>> type(gen)                  # but it returns a generator
<type 'generator'>
>>> hasattr(gen, '__iter__')   # that's an iterable
True
>>> hasattr(gen, 'next')       # and with .next (.__next__ in Python 3)
True                           # implements the iterator protocol.

The generator type is a sub-type of iterator:

>>> import collections, types
>>> issubclass(types.GeneratorType, collections.Iterator)
True

And if necessary, we can type-check like this:

>>> isinstance(gen, types.GeneratorType)
True
>>> isinstance(gen, collections.Iterator)
True

A feature of an Iterator is that once exhausted, you can't reuse or reset it:

>>> list(gen)
['I am', 'a generator!']
>>> list(gen)
[]

You'll have to make another if you want to use its functionality again (see footnote 2):

>>> list(func())
['I am', 'a generator!']

One can yield data programmatically, for example:

def func(an_iterable):
    for item in an_iterable:
        yield item

The above simple generator is also equivalent to the below - as of Python 3.3 (and not available in Python 2), you can use yield from:

def func(an_iterable):
    yield from an_iterable

However, yield from also allows for delegation to subgenerators, which will be explained in the following section on cooperative delegation with sub-coroutines.

Coroutines:

yield forms an expression that allows data to be sent into the generator (see footnote 3)

Here is an example, take note of the received variable, which will point to the data that is sent to the generator:

def bank_account(deposited, interest_rate):
    while True:
        calculated_interest = interest_rate * deposited 
        received = yield calculated_interest
        if received:
            deposited += received


>>> my_account = bank_account(1000, .05)

First, we must queue up the generator with the builtin function, next. It will call the appropriate next or __next__ method, depending on the version of Python you are using:

>>> first_year_interest = next(my_account)
>>> first_year_interest
50.0

And now we can send data into the generator. (Sending None is the same as calling next.) :

>>> next_year_interest = my_account.send(first_year_interest + 1000)
>>> next_year_interest
102.5

Cooperative Delegation to Sub-Coroutine with yield from

Now, recall that yield from is available in Python 3. This allows us to delegate coroutines to a subcoroutine:

def money_manager(expected_rate):
    # must receive deposited value from .send():
    under_management = yield                   # yield None to start.
    while True:
        try:
            additional_investment = yield expected_rate * under_management 
            if additional_investment:
                under_management += additional_investment
        except GeneratorExit:
            '''TODO: write function to send unclaimed funds to state'''
            raise
        finally:
            '''TODO: write function to mail tax info to client'''
        

def investment_account(deposited, manager):
    '''very simple model of an investment account that delegates to a manager'''
    # must queue up manager:
    next(manager)      # <- same as manager.send(None)
    # This is where we send the initial deposit to the manager:
    manager.send(deposited)
    try:
        yield from manager
    except GeneratorExit:
        return manager.close()  # delegate?

And now we can delegate functionality to a sub-generator and it can be used by a generator just as above:

my_manager = money_manager(.06)
my_account = investment_account(1000, my_manager)
first_year_return = next(my_account) # -> 60.0

Now simulate adding another 1,000 to the account plus the return on the account (60.0):

next_year_return = my_account.send(first_year_return + 1000)
next_year_return # 123.6

You can read more about the precise semantics of yield from in PEP 380.

Other Methods: close and throw

The close method raises GeneratorExit at the point the function execution was frozen. This will also be called by __del__ so you can put any cleanup code where you handle the GeneratorExit:

my_account.close()

You can also throw an exception which can be handled in the generator or propagated back to the user:

import sys
try:
    raise ValueError
except:
    my_manager.throw(*sys.exc_info())

Raises:

Traceback (most recent call last):
  File "<stdin>", line 4, in <module>
  File "<stdin>", line 6, in money_manager
  File "<stdin>", line 2, in <module>
ValueError

Conclusion

I believe I have covered all aspects of the following question:

What does the yield keyword do in Python?

It turns out that yield does a lot. I'm sure I could add even more thorough examples to this. If you want more or have some constructive criticism, let me know by commenting below.

Appendix:

Critique of the Top/Accepted Answer**

• It is confused on what makes an iterable, just using a list as an example. See my references above, but in summary: an iterable has an __iter__ method returning an iterator. An iterator provides a .next (Python 2 or .__next__ (Python 3) method, which is implicitly called by for loops until it raises StopIteration, and once it does, it will continue to do so.
• It then uses a generator expression to describe what a generator is. Since a generator is simply a convenient way to create an iterator, it only confuses the matter, and we still have not yet gotten to the yield part.
• In Controlling a generator exhaustion he calls the .next method, when instead he should use the builtin function, next. It would be an appropriate layer of indirection, because his code does not work in Python 3.
• Itertools? This was not relevant to what yield does at all.
• No discussion of the methods that yield provides along with the new functionality yield from in Python 3. The top/accepted answer is a very incomplete answer.

Critique of answer suggesting yield in a generator expression or comprehension.

The grammar currently allows any expression in a list comprehension.

expr_stmt: testlist_star_expr (annassign | augassign (yield_expr|testlist) |
                     ('=' (yield_expr|testlist_star_expr))*)
...
yield_expr: 'yield' [yield_arg]
yield_arg: 'from' test | testlist

Since yield is an expression, it has been touted by some as interesting to use it in comprehensions or generator expression - in spite of citing no particularly good use-case.

The CPython core developers are discussing deprecating its allowance. Here's a relevant post from the mailing list:

On 30 January 2017 at 19:05, Brett Cannon wrote:

On Sun, 29 Jan 2017 at 16:39 Craig Rodrigues wrote:

I'm OK with either approach. Leaving things the way they are in Python 3 is no good, IMHO.

My vote is it be a SyntaxError since you're not getting what you expect from the syntax.

I'd agree that's a sensible place for us to end up, as any code relying on the current behaviour is really too clever to be maintainable.

In terms of getting there, we'll likely want:

• SyntaxWarning or DeprecationWarning in 3.7
• Py3k warning in 2.7.x
• SyntaxError in 3.8

Cheers, Nick.

-- Nick Coghlan | ncoghlan at gmail.com | Brisbane, Australia

Further, there is an outstanding issue (10544) which seems to be pointing in the direction of this never being a good idea (PyPy, a Python implementation written in Python, is already raising syntax warnings.)

Bottom line, until the developers of CPython tell us otherwise: Don't put yield in a generator expression or comprehension.

The return statement in a generator

In Python 2:

In a generator function, the return statement is not allowed to include an expression_list. In that context, a bare return indicates that the generator is done and will cause StopIteration to be raised.

An expression_list is basically any number of expressions separated by commas - essentially, in Python 2, you can stop the generator with return, but you can't return a value.

In Python 3:

In a generator function, the return statement indicates that the generator is done and will cause StopIteration to be raised. The returned value (if any) is used as an argument to construct StopIteration and becomes the StopIteration.value attribute.

Footnotes

1. _{The languages CLU, Sather, and Icon were referenced in the proposal to introduce the concept of generators to Python. The general idea is that a function can maintain internal state and yield intermediate data points on demand by the user. This promised to be superior in performance to other approaches, including Python threading, which isn't even available on some systems.}

2. _{This means, for example, that range objects aren't Iterators, even though they are iterable, because they can be reused. Like lists, their __iter__ methods return iterator objects.}

yield was originally introduced as a statement, meaning that it could only appear at the beginning of a line in a code block. Now yield creates a yield expression. https://docs.python.org/2/reference/simple_stmts.html#grammar-token-yield_stmt This change was proposed to allow a user to send data into the generator just as one might receive it. To send data, one must be able to assign it to something, and for that, a statement just won't work.

First yield program

def countdown_gen(x):
      count = x
      while count > 0:
           yield count
           count -= 1

g = countdown_gen(5)

for item in g:
     print(item)

Output

5
4
3
2
1

Understand flow

Note

1. Return a generator obj Suspend a functon execution and save its
2. status, function can be executed again.

I was going to post "read page 19 of Beazley's 'Python: Essential Reference' for a quick description of generators", but so many others have posted good descriptions already.

Also, note that yield can be used in coroutines as the dual of their use in generator functions. Although it isn't the same use as your code snippet, (yield) can be used as an expression in a function. When a caller sends a value to the method using the send() method, then the coroutine will execute until the next (yield) statement is encountered.

Generators and coroutines are a cool way to set up data-flow type applications. I thought it would be worthwhile knowing about the other use of the yield statement in functions.

Yield is an object

A return in a function will return a single value.

If you want a function to return a huge set of values, use yield.

More importantly, yield is a barrier.

like barrier in the CUDA language, it will not transfer control until it gets completed.

That is, it will run the code in your function from the beginning until it hits yield. Then, it’ll return the first value of the loop.

Then, every other call will run the loop you have written in the function one more time, returning the next value until there isn't any value to return.

Yield

>>> def create_generator():
...    my_list = range(3)
...    for i in my_list:
...        yield i*i
...
>>> my_generator = create_generator() # create a generator
>>> print(my_generator) # my_generator is an object!
<generator object create_generator at 0xb7555c34>
>>> for i in my_generator:
...     print(i)
0
1
4

In short, you can see that the loop does not stop and continues to function even after the object or variable is sent (unlike return where the loop stops after execution).

yield yields something. It's like somebody asks you to make 5 cupcakes. If you are done with at least one cupcake, you can give it to them to eat while you make other cakes.

In [4]: def make_cake(numbers):
   ...:     for i in range(numbers):
   ...:         yield 'Cake {}'.format(i)
   ...:

In [5]: factory = make_cake(5)

Here factory is called a generator, which makes you cakes. If you call make_function, you get a generator instead of running that function. It is because when yield keyword is present in a function, it becomes a generator.

In [7]: next(factory)
Out[7]: 'Cake 0'

In [8]: next(factory)
Out[8]: 'Cake 1'

In [9]: next(factory)
Out[9]: 'Cake 2'

In [10]: next(factory)
Out[10]: 'Cake 3'

In [11]: next(factory)
Out[11]: 'Cake 4'

They consumed all the cakes, but they ask for one again.

In [12]: next(factory)
---------------------------------------------------------------------------
StopIteration                             Traceback (most recent call last)
<ipython-input-12-0f5c45da9774> in <module>
----> 1 next(factory)

StopIteration:

and they are being told to stop asking more. So once you consumed a generator you are done with it. You need to call make_cake again if you want more cakes. It is like placing another order for cupcakes.

In [13]: factory = make_cake(3)

In [14]: for cake in factory:
    ...:     print(cake)
    ...:
Cake 0
Cake 1
Cake 2

You can also use for loop with a generator like the one above.

One more example: Lets say you want a random password whenever you ask for it.

In [22]: import random

In [23]: import string

In [24]: def random_password_generator():
    ...:     while True:
    ...:         yield ''.join([random.choice(string.ascii_letters) for _ in range(8)])
    ...:

In [25]: rpg = random_password_generator()

In [26]: for i in range(3):
    ...:     print(next(rpg))
    ...:
FXpUBhhH
DdUDHoHn
dvtebEqG

In [27]: next(rpg)
Out[27]: 'mJbYRMNo'

Here rpg is a generator, which can generate an infinite number of random passwords. So we can also say that generators are useful when we don't know the length of the sequence, unlike list which has a finite number of elements.

Here's a simple yield based approach, to compute the fibonacci series, explained:

def fib(limit=50):
    a, b = 0, 1
    for i in range(limit):
       yield b
       a, b = b, a+b

When you enter this into your REPL and then try and call it, you'll get a mystifying result:

>>> fib()
<generator object fib at 0x7fa38394e3b8>

This is because the presence of yield signaled to Python that you want to create a generator, that is, an object that generates values on demand.

So, how do you generate these values? This can either be done directly by using the built-in function next, or, indirectly by feeding it to a construct that consumes values.

Using the built-in next() function, you directly invoke .next/__next__, forcing the generator to produce a value:

>>> g = fib()
>>> next(g)
1
>>> next(g)
1
>>> next(g)
2
>>> next(g)
3
>>> next(g)
5

Indirectly, if you provide fib to a for loop, a list initializer, a tuple initializer, or anything else that expects an object that generates/produces values, you'll "consume" the generator until no more values can be produced by it (and it returns):

results = []
for i in fib(30):       # consumes fib
    results.append(i) 
# can also be accomplished with
results = list(fib(30)) # consumes fib

Similarly, with a tuple initializer:

>>> tuple(fib(5))       # consumes fib
(1, 1, 2, 3, 5)

A generator differs from a function in the sense that it is lazy. It accomplishes this by maintaining it's local state and allowing you to resume whenever you need to.

When you first invoke fib by calling it:

f = fib()

Python compiles the function, encounters the yield keyword and simply returns a generator object back at you. Not very helpful it seems.

When you then request it generates the first value, directly or indirectly, it executes all statements that it finds, until it encounters a yield, it then yields back the value you supplied to yield and pauses. For an example that better demonstrates this, let's use some print calls (replace with print "text" if on Python 2):

def yielder(value):
    """ This is an infinite generator. Only use next on it """ 
    while 1:
        print("I'm going to generate the value for you")
        print("Then I'll pause for a while")
        yield value
        print("Let's go through it again.")

Now, enter in the REPL:

>>> gen = yielder("Hello, yield!")

you have a generator object now waiting for a command for it to generate a value. Use next and see what get's printed:

>>> next(gen) # runs until it finds a yield
I'm going to generate the value for you
Then I'll pause for a while
'Hello, yield!'

The unquoted results are what's printed. The quoted result is what is returned from yield. Call next again now:

>>> next(gen) # continues from yield and runs again
Let's go through it again.
I'm going to generate the value for you
Then I'll pause for a while
'Hello, yield!'

The generator remembers it was paused at yield value and resumes from there. The next message is printed and the search for the yield statement to pause at it performed again (due to the while loop).

Like every answer suggests, yield is used for creating a sequence generator. It's used for generating some sequence dynamically. For example, while reading a file line by line on a network, you can use the yield function as follows:

def getNextLines():
   while con.isOpen():
       yield con.read()

You can use it in your code as follows:

for line in getNextLines():
    doSomeThing(line)

Execution Control Transfer gotcha

The execution control will be transferred from getNextLines() to the for loop when yield is executed. Thus, every time getNextLines() is invoked, execution begins from the point where it was paused last time.

Thus in short, a function with the following code

def simpleYield():
    yield "first time"
    yield "second time"
    yield "third time"
    yield "Now some useful value {}".format(12)

for i in simpleYield():
    print i

will print

"first time"
"second time"
"third time"
"Now some useful value 12"

Shortcut to understanding yield

When you see a function with yield statements, apply this easy trick to understand what will happen:

1. Insert a line result = [] at the start of the function.
2. Replace each yield expr with result.append(expr).
3. Insert a line return result at the bottom of the function.
4. Yay - no more yield statements! Read and figure out code.
5. Compare function to the original definition.

This trick may give you an idea of the logic behind the function, but what actually happens with yield is significantly different than what happens in the list based approach. In many cases, the yield approach will be a lot more memory efficient and faster too. In other cases, this trick will get you stuck in an infinite loop, even though the original function works just fine. Read on to learn more...

Don't confuse your Iterables, Iterators, and Generators

First, the iterator protocol - when you write

for x in mylist:
    ...loop body...

Python performs the following two steps:

1. Gets an iterator for mylist:

   Call iter(mylist) -> this returns an object with a next() method (or __next__() in Python 3).

   [This is the step most people forget to tell you about]

2. Uses the iterator to loop over items:

   Keep calling the next() method on the iterator returned from step 1. The return value from next() is assigned to x and the loop body is executed. If an exception StopIteration is raised from within next(), it means there are no more values in the iterator and the loop is exited.

The truth is Python performs the above two steps anytime it wants to loop over the contents of an object - so it could be a for loop, but it could also be code like otherlist.extend(mylist) (where otherlist is a Python list).

Here mylist is an iterable because it implements the iterator protocol. In a user-defined class, you can implement the __iter__() method to make instances of your class iterable. This method should return an iterator. An iterator is an object with a next() method. It is possible to implement both __iter__() and next() on the same class, and have __iter__() return self. This will work for simple cases, but not when you want two iterators looping over the same object at the same time.

So that's the iterator protocol, many objects implement this protocol:

1. Built-in lists, dictionaries, tuples, sets, files.
2. User-defined classes that implement __iter__().
3. Generators.

Note that a for loop doesn't know what kind of object it's dealing with - it just follows the iterator protocol, and is happy to get item after item as it calls next(). Built-in lists return their items one by one, dictionaries return the keys one by one, files return the lines one by one, etc. And generators return... well that's where yield comes in:

def f123():
    yield 1
    yield 2
    yield 3

for item in f123():
    print item

Instead of yield statements, if you had three return statements in f123() only the first would get executed, and the function would exit. But f123() is no ordinary function. When f123() is called, it does not return any of the values in the yield statements! It returns a generator object. Also, the function does not really exit - it goes into a suspended state. When the for loop tries to loop over the generator object, the function resumes from its suspended state at the very next line after the yield it previously returned from, executes the next line of code, in this case, a yield statement, and returns that as the next item. This happens until the function exits, at which point the generator raises StopIteration, and the loop exits.

So the generator object is sort of like an adapter - at one end it exhibits the iterator protocol, by exposing __iter__() and next() methods to keep the for loop happy. At the other end, however, it runs the function just enough to get the next value out of it, and puts it back in suspended mode.

Why Use Generators?

Usually, you can write code that doesn't use generators but implements the same logic. One option is to use the temporary list 'trick' I mentioned before. That will not work in all cases, for e.g. if you have infinite loops, or it may make inefficient use of memory when you have a really long list. The other approach is to implement a new iterable class SomethingIter that keeps the state in instance members and performs the next logical step in it's next() (or __next__() in Python 3) method. Depending on the logic, the code inside the next() method may end up looking very complex and be prone to bugs. Here generators provide a clean and easy solution.

Yield is an object

A return in a function will return a single value.

If you want a function to return a huge set of values, use yield.

More importantly, yield is a barrier.

like barrier in the CUDA language, it will not transfer control until it gets completed.

That is, it will run the code in your function from the beginning until it hits yield. Then, it’ll return the first value of the loop.

Then, every other call will run the loop you have written in the function one more time, returning the next value until there isn't any value to return.

yield is just like return - it returns whatever you tell it to (as a generator). The difference is that the next time you call the generator, execution starts from the last call to the yield statement. Unlike return, the stack frame is not cleaned up when a yield occurs, however control is transferred back to the caller, so its state will resume the next time the function is called.

In the case of your code, the function get_child_candidates is acting like an iterator so that when you extend your list, it adds one element at a time to the new list.

list.extend calls an iterator until it's exhausted. In the case of the code sample you posted, it would be much clearer to just return a tuple and append that to the list.

All of the answers here are great; but only one of them (the most voted one) relates to how your code works. Others are relating to generators in general, and how they work.

So I won't repeat what generators are or what yields do; I think these are covered by great existing answers. However, after spending few hours trying to understand a similar code to yours, I'll break it down how it works.

Your code traverse a binary tree structure. Let's take this tree for example:

    5
   / \
  3   6
 / \   \
1   4   8

And another simpler implementation of a binary-search tree traversal:

class Node(object):
..
def __iter__(self):
    if self.has_left_child():
        for child in self.left:
            yield child

    yield self.val

    if self.has_right_child():
        for child in self.right:
            yield child

The execution code is on the Tree object, which implements __iter__ as this:

def __iter__(self):

    class EmptyIter():
        def next(self):
            raise StopIteration

    if self.root:
        return self.root.__iter__()
    return EmptyIter()

The while candidates statement can be replaced with for element in tree; Python translate this to

it = iter(TreeObj)  # returns iter(self.root) which calls self.root.__iter__()
for element in it: 
    .. process element .. 

Because Node.__iter__ function is a generator, the code inside it is executed per iteration. So the execution would look like this:

1. root element is first; check if it has left childs and for iterate them (let's call it it1 because its the first iterator object)
2. it has a child so the for is executed. The for child in self.left creates a new iterator from self.left, which is a Node object itself (it2)
3. Same logic as 2, and a new iterator is created (it3)
4. Now we reached the left end of the tree. it3 has no left childs so it continues and yield self.value
5. On the next call to next(it3) it raises StopIteration and exists since it has no right childs (it reaches to the end of the function without yield anything)
6. it1 and it2 are still active - they are not exhausted and calling next(it2) would yield values, not raise StopIteration
7. Now we are back to it2 context, and call next(it2) which continues where it stopped: right after the yield child statement. Since it has no more left childs it continues and yields it's self.val.

The catch here is that every iteration creates sub-iterators to traverse the tree, and holds the state of the current iterator. Once it reaches the end it traverse back the stack, and values are returned in the correct order (smallest yields value first).

Your code example did something similar in a different technique: it populated a one-element list for every child, then on the next iteration it pops it and run the function code on the current object (hence the self).

I hope this contributed a little to this legendary topic. I spent several good hours drawing this process to understand it.

yield is like a return element for a function. The difference is, that the yield element turns a function into a generator. A generator behaves just like a function until something is 'yielded'. The generator stops until it is next called, and continues from exactly the same point as it started. You can get a sequence of all the 'yielded' values in one, by calling list(generator()).

Here are some Python examples of how to actually implement generators as if Python did not provide syntactic sugar for them:

As a Python generator:

from itertools import islice

def fib_gen():
    a, b = 1, 1
    while True:
        yield a
        a, b = b, a + b

assert [1, 1, 2, 3, 5] == list(islice(fib_gen(), 5))

Using lexical closures instead of generators

def ftake(fnext, last):
    return [fnext() for _ in xrange(last)]

def fib_gen2():
    #funky scope due to python2.x workaround
    #for python 3.x use nonlocal
    def _():
        _.a, _.b = _.b, _.a + _.b
        return _.a
    _.a, _.b = 0, 1
    return _

assert [1,1,2,3,5] == ftake(fib_gen2(), 5)

Using object closures instead of generators (because ClosuresAndObjectsAreEquivalent)

class fib_gen3:
    def __init__(self):
        self.a, self.b = 1, 1

    def __call__(self):
        r = self.a
        self.a, self.b = self.b, self.a + self.b
        return r

assert [1,1,2,3,5] == ftake(fib_gen3(), 5)

Many people use return rather than yield, but in some cases yield can be more efficient and easier to work with.

Here is an example which yield is definitely best for:

return (in function)

import random

def return_dates():
    dates = [] # With 'return' you need to create a list then return it
    for i in range(5):
        date = random.choice(["1st", "2nd", "3rd", "4th", "5th", "6th", "7th", "8th", "9th", "10th"])
        dates.append(date)
    return dates

yield (in function)

def yield_dates():
    for i in range(5):
        date = random.choice(["1st", "2nd", "3rd", "4th", "5th", "6th", "7th", "8th", "9th", "10th"])
        yield date # 'yield' makes a generator automatically which works
                   # in a similar way. This is much more efficient.

Calling functions

dates_list = return_dates()
print(dates_list)
for i in dates_list:
    print(i)

dates_generator = yield_dates()
print(dates_generator)
for i in dates_generator:
    print(i)

Both functions do the same thing, but yield uses three lines instead of five and has one less variable to worry about.

This is the result from the code:

As you can see both functions do the same thing. The only difference is return_dates() gives a list and yield_dates() gives a generator.

A real life example would be something like reading a file line by line or if you just want to make a generator.

Can also send data back to the generator!

Indeed, as many answers here explain, using yield creates a generator.

You can use the yield keyword to send data back to a "live" generator.

Example:

Let's say we have a method which translates from english to some other language. And in the beginning of it, it does something which is heavy and should be done once. We want this method run forever (don't really know why.. :)), and receive words words to be translated.

def translator():
    # load all the words in English language and the translation to 'other lang'
    my_words_dict = {'hello': 'hello in other language', 'dog': 'dog in other language'}

    while True:
        word = (yield)
        yield my_words_dict.get(word, 'Unknown word...')

Running:

my_words_translator = translator()

next(my_words_translator)
print(my_words_translator.send('dog'))

next(my_words_translator)
print(my_words_translator.send('cat'))

will print:

dog in other language
Unknown word...

To summarise:

use send method inside a generator to send data back to the generator. To allow that, a (yield) is used.

An analogy could help to grasp the idea here:

Imagine that you have created an amazing machine that is capable of generating thousands and thousands of lightbulbs per day. The machine generates these lightbulbs in boxes with a unique serial number. You don't have enough space to store all these lightbulbs at the same time (i.e., you cannot keep up with the speed of the machine due to storage limitation), so you would like to adjust this machine to generate lightbulbs on demand.

Python generators don't differ much from this concept.

Imagine that you have a function x that generates unique serial numbers for the boxes. Obviously, you can have a very large number of such barcodes generated by the function. A wiser, and space efficient, option is to generate those serial numbers on-demand.

Machine's code:

def barcode_generator():
    serial_number = 10000  # Initial barcode
    while True:
        yield serial_number
        serial_number += 1


barcode = barcode_generator()
while True:
    number_of_lightbulbs_to_generate = int(input("How many lightbulbs to generate? "))
    barcodes = [next(barcode) for _ in range(number_of_lightbulbs_to_generate)]
    print(barcodes)

    # function_to_create_the_next_batch_of_lightbulbs(barcodes)

    produce_more = input("Produce more? [Y/n]: ")
    if produce_more == "n":
        break

As you can see we have a self-contained "function" to generate the next unique serial number each time. This function returns back a generator! As you can see we are not calling the function each time we need a new serial number, but we are using next() given the generator to obtain the next serial number.

Output:

How many lightbulbs to generate? 5
[10000, 10001, 10002, 10003, 10004]
Produce more? [Y/n]: y
How many lightbulbs to generate? 6
[10005, 10006, 10007, 10008, 10009, 10010]
Produce more? [Y/n]: y
How many lightbulbs to generate? 7
[10011, 10012, 10013, 10014, 10015, 10016, 10017]
Produce more? [Y/n]: n