Hi, some time ago I was looking for a function in itertools, actually, the combination functions allow us to have or not repetitions of element, have in consideration that if we allow repetitions there can be any amount of them, no way to limit the number of repetitions.

Here I asked in stackoverflow about this if there is a lib or option, in the end we wasn't able to found any about this and there was an answer with the solution.

I think this solution worth to be added to itertools, conceptually, the actual 2 functions of combinations are particular cases of this function.

https://stackoverflow.com/questions/67559105/combinatory-with-n-picked-elements-and-limited-repetitions-in-python/67559625

Well, there can be a extension of this too, where we can say, "how many times every particular element can be repeated".

Both of this things world be useful.

Thx.

How is proposed_function("abcd", max_repeat=3) any different from what you can currently spell as combinations("aaabbbcccddd") ? Or, more generally,

def proposed_function(it, repeats)
    repeated = chain.from_iterable([x] * repeat for x in it)
    return combinations(repeated)

This can easily generalize to specifying the max count of each item.

Dennis, combinations("aaabbbcccddd") isn't a valid call - the function requires a "how many?" argument too. If, e.g., we were asking for groups of 4, then combinations("aaabbbcccddd", 4) generates the 4-tuple ('a', 'b', 'c', 'd') 81 (3**4) times, while the OP presumably only wants to get it once.

OTOH, combinations_with_replacement("abcd", 4) can generate tuples with more than 3 repetitions of a given element.

The linked StackOverflow entry gives an efficient "real solution", but I agree with its author's final comment: "It is much too specialized for itertools". Indeed, it seems too obscure and special-purpose to me to even qualify as a reasonable candidate for an itertools doc "recipe".

> it seems too obscure and special-purpose to me to even
> qualify as a reasonable candidate for an itertools doc "recipe"

My thoughts are the same.

@latot Thank you for the suggestion but it doesn't make sense for the standard library.  We respectfully decline.