promptdojo_

Lists, dicts, and the shape of an API response — step 3 of 9

ch 03 · lists and dicts3/9

Indexing starts at zero, and -1 is the last item

This is the rule behind a lot of off-by-one bugs, and AI is not exempt. "Skipped the first item" and "crashed on the last row" often trace back to one thing: lists count from zero, and humans do not.

You'll trip over this once. Then it's automatic forever.

The two indexing rules

pets = ["luna", "moose", "biscuit"]
#         0       1         2       <- positive indexes
#        -3      -2        -1       <- negative indexes
  • Positive indexes count from the start, beginning at 0. The first item is pets[0]. The third item is pets[2].
  • Negative indexes count from the end, beginning at -1. The last item is pets[-1]. The second-to-last is pets[-2].

Read the editor on the right. With a 3-item list, pets[2] and pets[-1] happen to point at the same item — "biscuit" — because that's both "position 2 from the front" and "position 1 from the back". This is a coincidence of length, not a rule.

Why -1 is the move AI loves

When you ask Cursor for the last message, the most recent file, the latest entry, it almost always reaches for [-1]. Why? Because you don't have to know the length of the list. pets[len(pets) - 1] is the same as pets[-1], but the second one is shorter and length-agnostic.

You'll see this constantly in code that reads logs, tails feeds, processes the most-recent record. Any time the prompt has the words "last", "latest", "most recent", the code has a [-1] in it.

Where the off-by-one bug actually shows up

The classic AI flub: writing a for loop that walks a list and secretly skips the first item.

items = ["a", "b", "c"]
for i in range(1, len(items)):    # starts at 1, not 0!
    print(items[i])

That prints "b" and "c". The "a" got silently dropped because range(1, len(items)) starts at 1, but the first item is at index 0. Cursor sometimes writes this when you ask for "loop through the list," and the bug looks innocent until you notice your output is short by one.

The right version:

for i in range(len(items)):    # starts at 0
    print(items[i])

Or even better, the Python-native version:

for item in items:
    print(item)

We'll drill this exact pattern in the loops chapter. For now, when you see range(1, ...) in AI code, ask yourself: did it mean to skip index 0, or is that a bug?

What happens when you go off the end

Reading past the end of a list crashes:

pets[10]   # IndexError: list index out of range

There's no "default" or "empty string" fallback. Python raises an exception. It is a common runtime error in AI-generated code, and the fix is always the same: check the length first, or use a method that has a default (we'll cover dict.get later for the dict version).

Run the editor and watch the four prints. luna, moose, biscuit, biscuit — index 2 and index -1 line up in a 3-item list.

read, then continue.