Itertools

Itertools

What it is

itertools is a standard-library module of fast, memory-efficient iterator building blocks. itertools.chain(*iterables) is the most commonly reached-for one: it lazily iterates over several iterables in sequence, as if they were one, without first concatenating them into a new list.

Before this feature

list1 = [1, 2, 3]
list2 = [4, 5, 6]
list3 = [7, 8, 9]

combined = list1 + list2 + list3   # builds a whole new list in memory
for x in combined:
    process(x)

After this feature

from itertools import chain

list1 = [1, 2, 3]
list2 = [4, 5, 6]
list3 = [7, 8, 9]

for x in chain(list1, list2, list3):   # no intermediate list built
    process(x)

print(list(chain(list1, list2, list3)))  # [1, 2, 3, 4, 5, 6, 7, 8, 9]

# chain.from_iterable — for when you already have an iterable of iterables
matrix = [[1, 2], [3, 4], [5, 6]]
flat = list(chain.from_iterable(matrix))
print(flat)  # [1, 2, 3, 4, 5, 6]

Why this is better

• Lazy, not eager: chain() yields items one at a time as you iterate — no new combined list is materialized in memory, which matters when the inputs are large or are themselves generators.
• Works across mixed iterable types: chain a list, a range, a generator, and a set together without converting any of them first.
• Composable: itertools functions are designed to feed into each other — chain, islice, groupby, zip_longest, product, combinations — building complex iteration pipelines without writing manual loops.

Key notes / edge cases

• chain(a, b, c) takes the iterables as separate arguments; chain.from_iterable([a, b, c]) takes a single iterable of iterables — pick based on whether you already have a container of the inputs.
• Because chain() returns an iterator, you can only consume it once — list(chained) twice gives [] the second time.
• chain() doesn’t check for duplicates or sort anything — it’s purely sequential concatenation-by-iteration.

c = chain([1, 2], [3, 4])
print(list(c))  # [1, 2, 3, 4]
print(list(c))  # [] -- already exhausted

Quick practice

1. What does itertools.chain([1,2], [3,4], [5]) produce when converted to a list?

   Answer

   `[1, 2, 3, 4, 5]` — it iterates each input in sequence as if they were one iterable.

2. Why is chain() more memory-efficient than list1 + list2?

   Answer

   `chain()` yields items lazily as you iterate, without ever building a new combined list; `+` allocates a brand-new list holding every element up front.

3. When would you use chain.from_iterable instead of chain?

   Answer

   When you already have a single iterable *containing* the iterables to flatten (e.g. a list of lists), rather than the iterables as separate arguments.