Skip to content

python/front-insert-in-loop

Rule front-insert-in-loop
Pack Python
Severity warning
Fix none, edit by hand
Library versions every version

Each insert(0, x) moves every item already in the list, so a loop of them gets slow. Append in the loop and reverse once after it, or use a collections.deque.

def newest_first(events: list[Event]) -> list[Event]:
found: list[Event] = []
for event in events:
found.insert(0, event)
found.append(event)
found.reverse()
return found

The rule reports items.insert(0, item) in the body of a for or while loop, when the same function, or the module, starts items as [] or list() under a plain name. It leaves alone an insert outside a loop, an insert at another index, a list that the function takes as a parameter, an attribute, and a list that starts with items.

xpd has no automatic fix, so make the change by hand.

An insert at index 0 moves every item already in the list one place to the right, so the loop does work that grows with the square of the length of the list. On Python 3.15.2 on one machine, building a list of 10,000 items took 4.87 milliseconds with insert(0, item), 0.074 with append and one reverse() after the loop, and 0.103 with collections.deque and appendleft. The loop with insert took 1.39 times as long at 10 items, 2.06 times at 100 and 6.87 times at 1,000, and it was never faster. The three forms give equal contents.

Keep the deque when the loop also reads the front of the collection while it builds it, because the list is in the opposite order until reverse() runs. A deque is not a list, so call list(found) where the caller needs one, or slices it.

xpd prints this message with the issue. A name that starts with $ stands for code from your file.

Each insert at index 0 moves every item already in $L. Append in the loop and call $L.reverse() once after it, or start $L as collections.deque() and call $L.appendleft($X).

Browse the other rules of the Python pack.