near decade of algorithm solving, this is the first time i ever heard of 3ptr sliding window... cool
@jans3067
5 ай бұрын
Good solution! Two things I noticed: One, we can use an 'if' statement instead of a 'while' for the 'while len(count) < k' line since at any point in the array, the max length of count will be k+1. Two, l_near always points to an element with exactly one occurrence, so in that 'while' loop, we don't actually have to decrement it - we can always just pop it right away. So, we can just simplify the code with: if len(count) > k: count.pop(nums[l_near]) l_near += 1 l_far = l_near
@navyasaiporanki9450
4 ай бұрын
Yes, this is right! Good observation!!. I had the same question when I was going through a scenario and leftNear will always point to an element whose count is 1.
@laumatthew71
5 ай бұрын
Wow, sliding window technique on steroids... Great explanation, thanks!
@firstacc5442
5 ай бұрын
You improved sooo much compared to past tutorials, in these new tutorials, you including your thought process, which matches with all of our thought process at early stages of finding the solution for new problems and also explained why it won't work. Thank you soo much for your contribution to ever lasting best tutorials on youtube! Best of Luck!!
@oneepicsaxguy6067
5 ай бұрын
This can also be solved with one trick and simple 2 pointers You'll realise it is easier to find all subarrays with
@pranav7471
5 ай бұрын
Thats a nice trick! 2 times sliding window to find all sub arrays with
@upamanyumukharji3157
5 ай бұрын
You mean subtract
@chrischika7026
2 ай бұрын
@@upamanyumukharji3157 you are correct.
@prom3theus
5 ай бұрын
you are able to convey sliding 3 ptr window in the most intuitive way, thank you!
@torvasdh
5 ай бұрын
I almost solved this one on my own. Git to the 3 pointers idea but just couldnt work out the fine details. Ddnt help that it was like midnight lol Exciting because I think this was the first hard I almost solved on my own.
@vishaalagartha1658
5 ай бұрын
Nice! But I think a simpler way would be to use the 'At Most k, At Most k - 1' technique right?
@zziye810
5 ай бұрын
correct
@kapilkhandelwal48
5 ай бұрын
Yes it is more intuitive
@satyamjha68
5 ай бұрын
Yup
@MustafaAli-hr2vp
5 ай бұрын
true, but this is faster since it's one pass only
@De1n1ol
5 ай бұрын
@@MustafaAli-hr2vp it doesn’t matter. Both are O(n)
@RHR021397
5 ай бұрын
Thank you for your efforts to provide quality tutorials/solutions. Truly appreciated!
@itachid
5 ай бұрын
One thing to take care of is that the while loop which checks the length of the hash map should always come in first. I think that maybe this is because of the fact that the priority to check the length of the hash map is simply higher than the priority to check if we have more than one occurrence of a particular number.
@DroidHolicOfficial
5 ай бұрын
Another way to solve is like this - Subarrays with "EXACTLY" K Different Integers = Subarrays with 5 Subarrays So, subarrays with Exactly K different Integers = 12 - 5 => 7 So, we can iterate over the array twice, once to find number of subarrays with at most K different integers, and once to find the Number of subarrays with at most K - 1 different integers. And for both, we will use Sliding Window technique. In this way, the overall time complexity will be O(2N) => O(N) Here is my explanation on Leetcode for the same -> leetcode.com/problems/subarrays-with-k-different-integers/discuss/2582425/Python-Sliding-Window-%2B-Dictionary This is a technique we can use on other Sliding Window Problems as well which ask us to find subarrays with "EXACTLY" K something. For example this problem -> leetcode.com/problems/binary-subarrays-with-sum/ Or this problem -> leetcode.com/problems/count-number-of-nice-subarrays/
@user-rv1bx8hx4v
5 ай бұрын
Thank you! Great algorithm with 3 pointers 👍👍
@aayushtheapple
2 ай бұрын
C++ implementation : ``` class Solution { public: int subarraysWithKDistinct(vector& nums, int k) { unordered_map freq; int len = nums.size(), leftNear=0,leftFar=0, distinct=0, res=0; for(int right=0;rightk){ // left pointer will always point to the element whose frequency is 1. // while(distinct>k){ freq[nums[leftNear]]--; // if(!freq[nums[leftNear]]) // not needed as left will always point to the element whose frequency is 1. distinct--; leftNear++; leftFar = leftNear;// will update more times than we need. but it's okay } while(freq[nums[leftNear]]>1){ freq[nums[leftNear]]--; leftNear++; } if(distinct==k) res += leftNear-leftFar +1; } return res; } }; ```
@harshithdesai9989
5 ай бұрын
Great way of explaining the sliding window problem with three pointers.
@yang5843
5 ай бұрын
The fact that companies ask this in real interviews
@constantin1693
5 ай бұрын
Great job!! I've definitely heard about this 3 ptr technique, however forget. Thanks a lot for reminding!!
@rostislav_engineer
5 ай бұрын
Thank you for such video. With your explanation, this hard problem looks like an easy one!
@staywithmeforever
5 ай бұрын
i dont understand how people come up with getting no of sub arrays with two pointer like subtracting them we get subarrays I stuck in thinking but I know it works and know why does it work but still I can code myself in new problem for this too I thought 3 pointer but don't know how do make subarrays and over complicated taking 2 hashmaps and shifting farpointer and can even pass the base cases
@Anudeepindira
3 ай бұрын
In the scenario where len(count)> k, I see we aren't explicitly deleting all elements between l_far and l_near before the two pointers become equal. Is it implicitly being handled somehow? Ideally all the elements to the left of l_far should be made removed from hashmap.
@EduarteBDO
5 ай бұрын
After 4 hours I gave up on this problem and came here. After watching it I think that I would answer by myself if I spent a month.
@santanu29
5 ай бұрын
Great solution. I don't know how you come up with this.
@MustafaAli-hr2vp
5 ай бұрын
this explanation is way more intuitive than the editorial
@pastori2672
5 ай бұрын
all thanks to you i was capable of solving this problem using the hashmap last index trick and a sliding window
@pastori2672
5 ай бұрын
wow even the same 3ptrs technique
@erminiottone
5 ай бұрын
atMost(k) - atMost(k-1) is way easier to code and understand but I really appreciate this other solution because I learned a new pattern :)
@prabhas2445
5 ай бұрын
another easy way, atmost(k) - atmost(k-1) , atmost(k) can be easily done by sliding window
@Splish_Splash
5 ай бұрын
Thanks to you and hints I was able to solve today's hard question (2444)
@bahabouali6886
5 ай бұрын
the best leetcode videos you can find
@syedinayath4547
5 ай бұрын
Damn! Thanks for making the explanation so simple! I understood it in one go.
@johnniewalkerjohnniewalker2459
5 ай бұрын
interesting solution @NeetCodeIO.Well done!!!
@gryffindor6409
5 ай бұрын
hi neetcode, where r videos for todays contest and DCC?
@swanv951
5 ай бұрын
Is it true that for any subarray problems, an approach with window ending at a given index always works efficiently (and so, we never need to think about the alternative approach where the window starts at some index )?
@MP-ny3ep
5 ай бұрын
Beautiful explanation as always. Thank you
@michael._.
5 ай бұрын
damn I forgot about 3 pointers sliding window technique, gorgeous solution as always
@user-pv4xn3sg7j
5 ай бұрын
This was just wow man. Loved it 💯
@mhsunny123
5 ай бұрын
atMost trick neetcode did few videos ago for Leetcode 930. Binary Subarrays With Sum make implementation easy.
@get_out_it
5 ай бұрын
your skills are topnotch
@aaron-uz6pc
5 ай бұрын
Great explanation, thanks!
@satyamjha68
5 ай бұрын
Solved it !
@omkarjadhav6183
5 ай бұрын
I used to count subarry with less than equal to k and also less than equal to k -1 and then subtract the result to get the answer
@JAson-ps2ug
5 ай бұрын
great one pass solution, is (k) - (k-1) solution good enough for interview? I can only come up with that solution
for someone who codes in cpp here is the solution class Solution { public: int subarraysWithKDistinct(vector& nums, int k) { int n = nums.size(); int l_far = 0,l_near = 0,r = -1,ans = 0; unordered_map umap; while(++r < n){ umap[nums[r]]++; while(umap.size() > k){ umap[nums[l_near]]--; if(umap[nums[l_near]] == 0) umap.erase(nums[l_near]); l_near++; l_far = l_near; } while(umap[nums[l_near]] > 1){ umap[nums[l_near]]--; l_near++; } if(umap.size() == k){ ans = ans+(l_near-l_far+1); } } return ans; } };
@tarifahmed4956
5 ай бұрын
why you initialized r=-1?
@namanshah2688
5 ай бұрын
@@tarifahmed4956 i used it to make pre increment work in first while loop. U can also use r=0 and increment r at the end of first while loop
@theblogger4366
Ай бұрын
which app/website do you use for drawing?
@slizverg23
5 ай бұрын
Now that I’ve started to think that I begin to understand that “sliding window” thing, Neetcode says: “Ok, now we are gonna have two left pointers…”:)))
@Versatile_Naveen
5 ай бұрын
If we swap While loops inside a for loop it FAILSSSS why?????
@premranjan4440
5 ай бұрын
I just realised 2019 was 5 years ago!
@NeetCodeIO
5 ай бұрын
Yeah I'm slowly becoming a boomer. I still think of myself as a newgrad
@logchamption
5 ай бұрын
class Solution: def subarraysWithKDistinct(self, nums: List[int], k: int) -> int: def subA(nums,k): d = defaultdict(int) res = 0 l = 0 for r in range(len(nums)): d[nums[r]] += 1 while len(d) > k: d[nums[l]] -= 1 if d[nums[l]] == 0: del d[nums[l]] l += 1 res += r - l + 1 return res return subA(nums,k)-subA(nums,k-1)
@khatriiedits3606
5 ай бұрын
How did you come with the counting part?
@SC2Edu
5 ай бұрын
Wow, cool cool cool!
@ethanphelps5308
5 ай бұрын
anyone have a list of sliding 3 ptr window problems?
@ethanphelps5308
5 ай бұрын
Count Vowel Substrings of a String is a good one
@pjpodx
5 ай бұрын
did you write code in python on google interview ?
@NeetCodeIO
5 ай бұрын
yes
@CS_n00b
5 ай бұрын
Maybe I’m spending too much time doing leetcode…
@Munchen888
5 ай бұрын
NeetCode, pls explain 12:27 where from the result=6 is? Firstly, 1-0+1 = 2(6-2=4). Ok. Further 2 - 0 + 1 != 4 🤔 Other explanation is ok. But this moment I can’t catch …
@weihyac
5 ай бұрын
Position One: FAR.position = NEAR.position = 0 | 0-0+1=1 (count = 1) Position Two: FAR.position = 0, NEAR.position = 1 | 1-0+1=2 (incrementing count by 2 gives us 3) Position Three: FAR.position = 0, NEAR.position = 2 | 2-0+1=3 (incrementing count further by 3 gives us 6)
@Munchen888
5 ай бұрын
@@weihyac . Thank you for explanation.
@ks-xh4fq
5 ай бұрын
where is todays's video
@shreehari2589
5 ай бұрын
While else also works
@CTAAG-zp9nm
5 ай бұрын
class Solution { public int subarraysWithKDistinct(int[] nums, int k) { HashMap map = new HashMap(); int res = 0; int l_far = 0, l_near = 0, r = 0; while (r < nums.length) { map.put(nums[r], map.getOrDefault(nums[r], 0) + 1); while (map.size() > k) { map.put(nums[r], map.getOrDefault(nums[r], 0) - 1); if (map.get(nums[r]) == 0) { map.remove(nums[r]); } if(l_near < nums.length ) l_near+=1; l_far = l_near; } while ( map.get(nums[l_near]) != null || map.get(nums[l_near]) > 1) { map.put(nums[r], map.getOrDefault(nums[r], 0) - 1); if(l_near < nums.length ) l_near+=1; } if (map.size() == k) { res += (l_near - l_far + 1 ); } r++; } return res; } } i am getting error can any 1 point out why and how to solvey them
Пікірлер: 81