Super "fun" #Python bug: groupby in itertools ostensibly groups items by the result of a key function, however it just groups similar sequential items in the input list. You can't safely use it directly to build a dictionary. It will silently drop some of your items.
>>> data = ['aa', 'bbb', 'cc']
>>> grouped = {k: list(v) for k, v in groupby(data, key=len)}
>>> print(grouped)
{2: ['cc'], 3: ['bbb']}
I remember misunderstanding the niche use of groupby years ago because the documentation is a bit subtle on the fact that it groups consecutive elements.
@treyhunner@slott56 It's a function naming bug. "groupby" means something very specific that this function does not do.
Ask 100 engineers what this function does and 95 of them will give you the sql "group by" behavior answer, while the other 5 will be people who've been burned by this.
I named a #Python exercise "group_by" many years ago because it's an apt description of what it does even though itertools.groupby won't be of much help for solving it.
Add comment