There is something very strange I saw in this last contest "Rayan"

Why many LGMs took time like 10-20 mins to solve div2D (this time is equivalent to 2 or 3 hrs of me or maybe more :) ), when I saw that and it was the end of the contest I stopped thinking automatically (not letteraly, like with no motivation) then suddenly many of my friends some of them are Expert about 1750 solved it in nearly this average time and another friend who is ~1150 solved it in nearly less than hr (but unofficial)

is this because constructive problems normally take time or what?

UPD2: It is my first time to make a post here, can I know why many people are making downvotes (nearly those didn't write a comment), I think if they told me the reason for that maybe this will be helpful for both of us

Again, someone hacked my B solution coz of the Python Dictionary (this can happen also when using CPP unordered_map)

But I noticed that it can be done using an array because the constraint bi <= k which will allow to avoid Py dict

So, I wonder did the problem setter put this constraint for Python users to be able to solve it also in Python as they can avoid dictionary because of it, or the constraint exists for any other reason?

UPD1: Now, I have used d[str(key)] = value and it works and got ACC not TLE like the previous

Submission: 289783097

brands = defaultdict(int)
for _ in range(k):
    bi, ci = read_numbers()
    brands[str(bi)] += ci

brands = list(brands.values())
brands.sort(reverse=True)
ans = 0
for i in range(min(n, len(brands))):
    ans += brands[i]

print(ans)

the only change is the str(bi) instead of bi, can anyone try to hack it or show me something that will let this not work

coz I really wants to use Python all time (recursion problem has no idea how to solve it yet, but I am speaking now about this problem)