[Tutor] Finding the shortest word in a list of words

[John Fouhy]

2009/1/20 Emad Nawfal (عماد نوفل) <emadnawfal at gmail.com>:
> Hello tutors,
> I need to find the shortest / longest word(s) in a sequence of words. I've
> done this and it works, but I'm wondering whether this is a good way:
>>>> words = "man woman children he".split()
>>>> words
> ['man', 'woman', 'children', 'he']
>>>> lens = [len(word) for word in words]
>>>> lens
> [3, 5, 8, 2]
>>>> for word in words:
> ...     if len(word) == min(lens): print word
> ...
> he

Hi Emad,

You can use the decorate-sort-undecorate idiom to make this technique
a bit nicer.

"decorate" means "add information to the things in the list":

>>> words_and_lengths = [(len(w), w) for w in words]
>>> words_and_lengths
[(3, 'man'), (5, 'woman'), (8, 'children'), (2, 'he')]

Now I can sort it and get the shortest element easily:

>>> words_and_lengths.sort()
>>> words_and_lengths[0]
(2, 'he')

Or I can undecorate:

>>> words2 = [w[1] for w in words_and_lengths]
>>> words2
['he', 'man', 'woman', 'children']

Python 2.5+ provides another way of doing this, using the key=
argument to sort():

>>> words
['man', 'woman', 'children', 'he']
>>> words.sort(key=len)
>>> words
['he', 'man', 'woman', 'children']

This essentially does the decorate-sort-undecorate in one step, where
len is the function we used to do the decoration.

Of course, this is not necessarily the best answer for your particular
problem.  The problem with sorting is that you have to look at some
elements more than once.  For short lists, it's not a problem, but it
can slow you down on bigger lists.  You could also find the shortest
element by going through the list, remembering the shortest element
you've seen so far.  This will be quicker if you only want to find the
single shortest.

-- 
John.