Kacam.net

Algorithms

求第 k 大,为什么用小顶堆

求最大的 100 个,用小顶堆 —— 方向是反的,为什么?

Copy the prompt, then pick a model. It plays the interviewer and asks you through it scene by scene instead of handing you the answer.

On a phone with the app installed, the system opens the app directly. Links cannot open external apps inside WeChat — open it in a browser first.

The copied text ends with the answer key the model grades you against. Don't read it yourself first — working it out on your own is the point.

「Top-K 用堆」人人会答,但「大顶堆还是小顶堆」这一问就能分出会和背。正确的心智模型是:堆顶是门槛——目前入选的 k 个里最差的那个,新来的比它好就顶掉它。

这一份还会逼你先问清楚题(数据在内存还是流式、内存多大、要前 k 个还是第 k 个),上来就说「用堆」的会被打断。10 亿个 int 是 4 GB,空间才是这题的重点。

★★★☆☆ · about 30 min · Tests: 堆顶作为门槛、建堆的真实复杂度、什么时候不该用堆

What this one drills

Choosing a structure Complexity Cache and constants

Have you done this one?

Sign in to mark this one as practised — next time you come back you'll see when you did it.

Notes from people who did this one

Nobody has written yet. After you run it, say what came out of the conversation — which scene stumped you, which conclusion you changed afterwards.

Sign in to write a note.

How to use this

This set teaches rather than interrogates: it guides you instead of pressing until you have nothing left to say. When you're stuck it first asks which layer you're stuck at (never heard of it / know it but not why / know why but can't write it), then decides what to give. Only after guiding you twice on the same point will it spell the answer out — and then it asks you to say it back in your own words.

How to start

  1. Pick one, open it, copy the prompt.
  2. Paste it into a brand-new conversation, say nothing, and wait for its first question.
  3. Trace by hand before you write code — it will interrupt you. Walking through an array of length 2 beats writing the template ten times.

Four commands you can call out at any time

  • 跳过 move to the next scene now
  • 换个说法 explain this bit again from another angle
  • 直说吧 spell the answer out, then make me say it back
  • 复盘 jump to the wrap-up and print the review card

Three traps

  • Don't read the answer key at the end first — that "how people get it wrong" section is the answer to every question it will ask you.
  • Don't rush to code. With algorithms, writing code is the easiest way to hide the fact that you haven't actually thought it through.
  • When it starts praising you, point it back to the "don't praise me" rule. Guiding is not coddling.

Can't come up with a solution? Start with defining the DP state. Always off by one? Binary search and prefix sums. Can't recognise the shape of a problem? Monotonic stack and sliding window. The complexity one underpins several others — do it early.