Chunk with batched, compare neighbours with pairwise

Two jobs come up constantly and are both easy to get off by one: splitting a list into chunks of N, and comparing each item with the one before it. itertools has a named function for each, batched (Python 3.12+) and pairwise (3.10+).

Before

ids = list(range(1, 11))
for i in range(0, len(ids), 4):
    print(ids[i:i + 4])

readings = [12.0, 15.5, 15.0, 19.25]
for i in range(1, len(readings)):
    print(readings[i] - readings[i - 1])

[1, 2, 3, 4]
[5, 6, 7, 8]
[9, 10]
3.5
-0.5
4.25

It works, but every index is a place for len - 1 or i + 1 to go wrong, and it only works on things you can index.

After

from itertools import batched, pairwise

ids = range(1, 11)
for chunk in batched(ids, 4):
    print(chunk)

readings = [12.0, 15.5, 15.0, 19.25]
for before, after in pairwise(readings):
    print(after - before)

(1, 2, 3, 4)
(5, 6, 7, 8)
(9, 10)
3.5
-0.5
4.25

The chunks come out as tuples, and the last one is simply shorter. Both work on any iterable, including a file or a generator, with no indexing at all.

And one more: flatten one level

chain.from_iterable joins a list of lists without a nested loop or sum(lists, []).

from itertools import chain

weekly = [["Mon", "Tue"], ["Wed"], ["Thu", "Fri"]]
print(list(chain.from_iterable(weekly)))

['Mon', 'Tue', 'Wed', 'Thu', 'Fri']

Why it works

Each is a small iterator that remembers its place. batched pulls N items at a time; pairwise keeps the previous item and hands you (previous, current). Neither copies your data first, so they work the same on ten items or ten million.

When not to use it

batched is Python 3.12 or newer, so older servers need the slicing loop (the browser here runs 3.13). If you need the chunks padded to equal length, that's a different recipe (zip_longest), and for rolling windows wider than two, a collections.deque(maxlen=N) reads more clearly.