Code Review Stack Exchange is a question and answer site for peer programmer code reviews. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Pythonic way of expressing the simple problem:

Tell if the list needle is sublist of haystack

#!/usr/bin/env python3

def sublist (haystack, needle):

    def start ():
        i = iter(needle)
        return next(i), i

        n0, i = start()
        for h in haystack:
            if h == n0:
                n0 = next(i)
                n0, i = start()
    except StopIteration:
        return True
    return False
share|improve this question
So, what's the question? – Jeff Vanzella Dec 15 '12 at 3:27
@JeffVanzella: Fair point. That's "code review". The question is: "what do you think in general?". – Dacav Dec 15 '12 at 12:30
up vote 5 down vote accepted

1. Bug

Here's a case where your function fails:

>>> sublist([1, 1, 2], [1, 2])

this is because in the else: case you go back to the beginning of the needle, but you keep going forward in the haystack, possibly skipping over a match. In the test case, your function tries matching with the alignment shown below, which fails at the position marked X:

haystack   1,1,2
needle     1,2

Then it starts over from the beginning of the needle, but keeps going forward in the haystack, thus missing the match:

haystack   1,1,2
needle         1,2

So after a mismatch you need to go backward an appropriate distance in the haystack before starting over again from the beginning of the needle.

2. A better algorithm

It turns out to be better to start matching from the end of the needle. If this fails to match, we can skip forward several steps: possibly the whole length of the needle. For example, in this situation:

haystack  1,2,3,4,6,1,2,3,4,5
needle    1,2,3,4,5

we can skip forward by the whole length of the needle (because 6 does not appear in the needle). The next alignment we need to try is this:

                    O O O O O
haystack  1,2,3,4,6,1,2,3,4,5
needle              1,2,3,4,5

However, we can't always skip forward the whole length of the needle. The distance we can skip depends on the item that fails to match. In this situation:

haystack  1,2,3,4,1,2,3,4,5
needle    1,2,3,4,5

we should skip forward by 4, to bring the 1s into alignment.

Making these ideas precise leads to the Boyer–Moore–Horspool algorithm:

def find(haystack, needle):
    """Return the index at which the sequence needle appears in the
    sequence haystack, or -1 if it is not found, using the Boyer-
    Moore-Horspool algorithm. The elements of needle and haystack must
    be hashable.

    >>> find([1, 1, 2], [1, 2])

    h = len(haystack)
    n = len(needle)
    skip = {needle[i]: n - i - 1 for i in range(n - 1)}
    i = n - 1
    while i < h:
        for j in range(n):
            if haystack[i - j] != needle[-j - 1]:
                i += skip.get(haystack[i], n)
            return i - n + 1
    return -1
share|improve this answer
@Lattyware: Generally I'd agree that you should use iterators wherever it makes sense, but this is one of the exceptional cases where iterators don't make much sense. A pure-iterator search runs in O(nm) whereas Boyer–Moore-Horspool is O(n). (In the average case on random text.) Sometimes you just have to get your hands grubby. – Gareth Rees Dec 15 '12 at 2:11
Scratch all my arguments, it's impossible to do this correctly the way I was doing it. – Latty Dec 15 '12 at 2:29
Accepted. The source is strong with this one. – Dacav Dec 15 '12 at 12:38
Does the dict comprehension overwrite previous values? If not you'll skip too far when you see e while searching for seven. I'd bet it does but am not sure. I do miss Python for its elegance. – David Harkness Feb 6 '14 at 3:14
Yes, later items in the dict comprehension override previous items, and so find('seven', 'even')1 as required. – Gareth Rees Feb 6 '14 at 10:17

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.