Search code examples
pythonthread-safetytee

Is the result of itertools.tee() thread-safe (Python)

Suppose I have this Python code:

from itertools import count, tee
original = count()     # just an example, can be another iterable
a, b = tee(original)

The question is, will there be any problem if I start iterating "a" in one thread and, at the same time, iterating "b" in another thread? Clearly, a and b share some data (the original iterable, + some additional stuff, internal buffers or something). So, will a.next() and b.next() do the appropriate locking when they access this shared data?

Solution

  • In C-Python, itertools.tee() and the iterator it returns are implemented with C code. That means the GIL should protected it against being called simultaneously by multiple threads. It will probably work correctly, and it won't crash the interpreter, but it is not guaranteed to be thread safe.

    Simply put, don't take the risk.