Generator problem in code playground

Question

I have a code (random int generator) that works fine in IDLE but behaves strangely in code playground. Here on Sololearn it seems to yield the first number over and over. What is wrong with it?

https://code.sololearn.com/cE6saubSp1AR/#py

Code playground output:
10 random numbers in range(0, 100):
4
4
4
4
4
4
4
4
4
4

IDLE output:
10 random numbers in range(0, 100):
15
35
15
47
27
19
39
3
95
47

Answer 1

Checking infinite. Edit:

def natural(stopat):
i = 1
while i<=stopat:
yield i
i += 1

def infnat():
i=1
while True:
yield i
i += 1

if __name__ == '__main__':
for i in natural(10):
print(i)

infgen=infnat() # one generator object
for i in range(10):
print(next(infgen)) # reuse object

Outputs:
1..10 (as limited sequence)
1..10 (as infinite sequence)

====
Your code - move rand(a,b) out as r:

def test_rand(numtests, a, b):
print...
r=rand(a,b)
for i in range(numtests):
print(next(r))

Answer 2

edit...

http://stackoverflow.com/a/34344919/3981745

Recommends using a UUID. I'm poking at your code, which is returning non-unique times.

Answer 3

update 1: ...but that's what stalled me. The function is supposed to resume execution after the last yield(), but when I insert print(seed) after its initialization (print() at line 11), I see the same hash (and actually the same timestamp) 10 times.

This made me think I didn't understand the generator. I'll look again.

Answer 4

Update 2: I bumped your request from 10 to 1000. About halfway through the output changed, then got stuck on that number.

Answer 5

Not the hash...I'm saying it should change which implies you're potentially reinitializing your generator.

Using your 'natural' example:

def natural(stopat):
i = 1
while i<=stopat:
yield i
i += 1

if __name__ == '__main__':
for i in natural(10):
print(i)

Answer 6

It appears that to make a generator object, you have to assign the generating function to a variable. Otherwise codeplaygroung considers it as just a function and calls it from the very beginning each time:

def natural():
i = 1
while True:
yield i
i += 1

if __name__ == '__main__':
nat = natural()
print(type(natural))
print(type(nat))

Output:
<class 'function'>
<class 'generator'>

Answer 7

Thanks, @Kirk. But the problem has nothing to do with hash(time()). I use it only once in my program as a seed value. Take a look here: https://code.sololearn.com/cx6A4eVw6pNw/#py

def natural():
i = 1
while True:
yield i
i += 1

if __name__ == '__main__':
for i in range(10):
print(next(natural()))

There is no randomness in the code. But it still yields only the first element.

Answer 8

Update 3:

I tried .__next__() and .next() - both problems.

But generators maintain state...so I gave the generator limit control by passing numtests and then 'while True' to 'while c < alltests', and changing the request loop from range() to 'in'.

This resulted in varying hashes.

c=0
while c<alltests:
...
c+=1
yield ...

... and below:

for i in rand(a,b,numtests):
print(i)

Answer 9

Just wow @Kirk. Good job. You made it work. Cheers!

Answer 10

yeah, that's weird... Just contact SoloLearn to report the bug, it's quite peculiar...

Answer 11

Yes, limited iterators work fine here. But what if I want an infinite one? And how can i reinitialize a generator unawares?