Yield is for generator functions. Return ends the function and sends back the specified value.
Generators are functions that return iterators, i.e. they 'yield' a value after each iteration of them, before proceeding to the next one. This is generally used for things where you don't need all the results returned at once (which must load them all into memory), but instead only gives back the results one at a time (which can vastly reduce memory usage).
# start at the last index and step back one each time.
for index in range(len(data)-1, -1, -1):
for char in reverse('golf'):
What this does is call each iteration of the reverse function each time the "for" loop executes in the statement for char in reverse('golf') . This way, only a single letter is sent back each time, rather than the fully-reversed string.
This isn't a great example of generator usage, but say if you had a dataset with 1000s of numbers and you needed to perform calculations on them, but you'd only be working on one item in the set at a time anyway. Then a generator would be useful as it would save a TON of memory by only returning each item as it's needed.
Hope that helped!
The point is so that you can keep whatever changes you've made to a value. Like:
# double-up a string
return s + s
a = "Hello world!"
a = doubleUp(a)
b = doubleUp(a)
print a # should display Hello World!Hello World!
print b # should display Hello World!Hello World!Hello World!Hello World!
If I just said s = s + s in the function, 'a' wouldn't be changed by saying doubleUp(a) as 's' is a variable in the local scope of the function. If I did this:
s = s + s
a = "Hello World!"
print a # should display Hello World!
As you can see in the second example, all we did was change the value of 's' inside the function, but a isn't changed because we did not return the value of 's' and assign it back to 'a'.
In C++ you could pass either by pointer or by value, but Python's way of passing is a bit more complicated...
Hope I clarified that!