in the last codeforces contest Codeforces Round 950 (Div 3) , i made up to 500 hacks for both solutions of c++ and python. but why?
First let me explain why for Python : the main collections that hack the sol. in python is Counter (collections) , but why?
It is not an issue related to python slowness, but it is related to hash collisions being o(n^2) but imagine with me if you : sort the array before you use counter, it should pass or try random.shuffle on points instead of sorting.
why the sorting make it pass ?
People have designed the testcase to cause hash collision resulting in o(n^2). If you sort the array you get rid of the hash collision as elements are inserted in a different order, resulting in o(n).
Second and the usual, why for Cpp :
check the blog to make it o(n) instead of o(n^2) : https://mirror.codeforces.com/blog/entry/62393.
and Finally, i hope that i made you benefit.
thanks
interesting blog
Thx for your interest in clarifying this
Hi, today I solved these two problems:
First, I solved them using a
map
. Then, out of curiosity (stemming from what happened in the last round), I submitted the same solutions but using anunordered_map
and still got accepted for both problems.Is this because there are no anti hash tests? I’m still unsure when to rely on the use of an
unordered_map
, unless one implement a pretty good hash function, are those solutions always hackable?That was helpful. Thanks!
hope to benefit you :)