Grind75

Tasking this out now, as well as timing it.

542, 973

What's tough here is the acceptance criteria. It needs to be 3 unique indices and the combination itself needs to be unique. Having the nested 3-loop is cleaner because the variables are in scope. Doing it with the backtracking approach requires cramming data in the recursive function.

dead end. Moving on

The loops are moved around and hidden, but there still seems to be 3 distinct loops. The thing that's interesting (ish) here is that they are treating the problem as an extension of twosum, and re-using that logic, calling a version of twosum inside of 3sum.

I like their algorithm, which looks the two-pointer in the editorial.

I hate when people talk in the library. I can't focus.

This graph problem isn't clicking for me yet (LC 207). Grr. Taking a break.

Attempting that coin change problem again, among other things.

Per the recommendation of several people, I am going to just the Leetcode website to write code. It has the verification and eveything, and will help me pick up on weird edge cases.

39 (combination sum) is a backtracking problem that I am not understanding properly.

I did 46, poorly. 39 passed when I used a different data structure.

Uh oh, I've just been running locally. I've been forgetting to submit stuff. I knew things were feeling wrong.

56, 981

Even with the editorial in my face, it still took me an hour to implement. jeez.

Pen and paper

Coding it up

Only took me 2.5 hours.

Just want to get that last problem working (LC 310: minimum height trees)

I'm trying to grok this findMinHeightTrees problem better. Also, LC 739 (Daily Temperatures)

198: House Robber

134: Gas Station

134: Gas Station (Continued) #grind75

More Gas Station (134)

Gas Station, again (134). I thought it was dynamic programming. It was not. Another day wasted. I will have to look at this: https://www.youtube.com/watch?v=lJwbPZGo05A.

Watching the neetcode video.

36: Valid Sudoku

49: Group anagrams

152: maximum product subarray

211: Design and Search Words Data Structure. I think it's a Trie problem. Spent most of this time just getting the Trie structure the way I imagined it. Have not implemented the search yet, which I believe will be a DFS using a stack (for backtracking when a '.' is encountered).

Ran out of time. Right data struture (trie), but my search doesn't work properly ('..' is a problem on the huge edge case).

I'm breaking out too early.

Nevermind. I'm not getting it right. Time to stop.

Ah. I was not making a Trie in my original implementation. I only kept track of possible combinations for each position, which was more efficience memory-wise, but it did not keep track of paths between unique words.

Example: "dig" and "dog" show up as (d (o (g)) (i (g))) There are two g's here. My structure represented the set as ([d], [o i], [g]).

If I were to add "mop" to the the structure. A trie would look like: ((m (o (p)) (d (o (g)) (i (g)))) My structure would look like: ([m d] [o i] [g p]) There's no way to tell if "mip" is a valid path or not

Even with pointers, "mog" could be valid with this path. ([m:[o] d:[o i]] [o:[g p] i:[g]] [g p])

417: Pacific Island Water Problem. I think I am close. It seems to lend itself well to a dynamic programming problem, but I turned up an edge case where I wasn't checking in all the available directions. Will look into that later.

I'm hoping I just need to code up an edge case and that my overall solution is solid.

My edge case isn't working.

If I do 6 months, it'll be cheaper than the monthly rate.

Remove Nth node from linked list

Shortest path to get food. Note that the time here is for implementation. Took me about 3-5 extra minutes to get the implementation, but I casually timed that one.

Find the duplicate number. Did brute force, then looked up the answer.

Lexicographic order threw me off. It turns out the built-in structures in Python used to solve this problem (Counter, heapify) seem to handle this for you already? Was I supposed to know that?

Follow-up problem 347 Top K frequent elements. Apparently this follows a similar pattern: https://leetcode.com/problems/top-k-frequent-elements/description/

I did this already I think. But I didn't submit it, so it's a redo now.

Note to self: a subarray is different from a subsequence. Subarrays are continguous.

Wow, I did not get this right the first time. I gave up early and looked at the answer.

I didn't submit. So, I didn't really learn anything. yay.

Got something working, but I want to review this tomorrow

Initial attmpts. Program hit edge cases quickly.

I was close to getting BFS right. I didn't think to simply erase islands as I went.