Iterators and lazy pipelines

The idea

Iteration is a protocol: an iterable can provide an iterator, and an iterator supplies successive values until it is exhausted. A list can produce a fresh iterator repeatedly, while a particular generator object tracks one position. A function containing yield creates a generator when called; its body runs as values are requested rather than computing every result immediately.

Generators are useful when a pipeline can process one item at a time or stop early. They reduce retained intermediate data, but they do not make an expensive operation inherently cheap. Once consumed, a generator does not rewind. If two independent consumers need the complete sequence, recreate it from a repeatable source or deliberately materialize it, considering the memory cost.

def squares(values):
    for value in values:
        yield value * value

items = squares([2, 3])
print(next(items))
print(list(items))
print(list(items))

4
[9]
[]

next consumes the first yielded value. The first list call consumes the remainder; the second sees an exhausted generator.