Source: 📖 Effective Python item 30

Using generators

A generator is produced by a function that utilises the yield expression. When a generator function is called, it doesn't actually run—instead it returns an iterator. The next function can then be used on this new iterator, causing the generator to advance to the next yield expression. Each value passed to the yield expression is then returned by the generator function to the caller. This means that generators have better memory management than returning a full list, as returning a full list requires the full list to be created before it can be returned. This can use up a system's available memory when working with large list outputs. By contrast, a generator only deals with one value at a time, keeping memory consumption at safe levels.

def fib_generator():
	a = 0
	b = 1
	c = 0
	n = 1
	if n == 1:
		n += 1
		c = b + a
		a = b
		b = c
		yield 0
	elif n == 2:
		n += 1
		c = b + a
		a = b
		b = c
		yield 1
	else:
		n += 1
		c = b + a
		a = b
		b = c
		yield c

fib_iterator = fib_generator()

for _ in range(8):
	print(next(fib_iterator))

>>>
0
1
1
2
3
5
8
13

The above generator can be used to retrieve one Fibonacci number at a time each time the next function is called on fib_iterator.