ByteDance program yuan teaches you how to brush algorithm questions: I'm not afraid of the interviewer tearing the code

Write it at the front

“ The sea of questions is vast ,leetcode Thousands of . How should I brush the question ？”
“ The title is finished , During the interview, I tore the code by hand or my head was empty , What do I do ？”
You are preparing for spring 、 Autumn recruit interview algorithm question is not confused by these two questions ？

So congratulations , You found this treasure article . This article will show you a decisive guide for brushing questions . And from the importance of Algorithm 、 The selection of question bank and the three steps of writing questions are explained separately , Completely solve all your problems about the interview Algorithm .

reminder , This article will not teach you how to improve the level of question brushing , I didn't run to the Blue Bridge Cup 、ACM And other record breaking standard sprints in various algorithm competitions . The author himself is a programmer with mediocre IQ , There is nothing to say about the export of competition experience , I was a freshman last time C Language final exam …… But the algorithm assessment in the interview has been maintained until the time of landing 100% A glorious record of accuracy .

The purpose of this article is pure and true ： Teach you to brush the least questions , In the least amount of time , The safest way to win the interview algorithm question ！

If your code ability is mediocre, novice Xiaobai , Then this article will definitely benefit you a lot ; If you have won various gold and silver awards , Then be the front line of the network , Meeting is fate . Look at this fate , I'll go after you like it ～

Okay , That's bullshit .

Come on , Let's start ！

Algorithm , As one of the most important assessment standards for school recruitment and even social recruitment , It is also the key evidence for passing the assessment . The candidates are preparing for the interview “ highlight ”.

First , Do you know why hand tearing code is so important ？

Because the algorithm ability can show the candidate's basic computer ability . The basic capabilities here include ： Solid programming skills 、 Good code habits and problem-solving thinking ability .

You may be dismissive , Is the algorithm really useful in practical work ？

To be honest , Very useful and very “ useless ”.
It's useless after you really enter the job , Busy business and tight demand may make you have no time to think about how to apply various algorithms , Directly a lot of reusable code copy; But because of these “ Without thinking ” The code is overrun , One for code rigor 、 High standard of robustness 、 Strict requirements of talent has become the interviewer's eyes “ The bestselling ”.

One can save this pile “ Junk code ” People are simply God ！

therefore , Algorithmic capabilities for long-term requirements iteration and stability of Engineering Architecture , It's very, very important , It is also the ability that programmers must have to move bricks and yards from ordinary farmers to architects .
Since it's so important , As the core evaluation standard of the interview , Not too much ？

Next , You may be curious again ： Yien sister , Why did you say , This article cannot improve the algorithm ability , But I can tear down the code ？
Good question. , Let me tell you ：

What is the law of the interviewer tearing the code , What's the difference with general algorithm problems ？

First of all, let me share with you a treasure website ：https://github.com/afatcoder/LeetcodeTop,github Of star Broken 12k（2022.02）, It is the most comprehensive and best interview algorithm I have ever seen .

Its title comes from Niuke net's surface Sutra and brush Sutra 、CSDN And enthusiastic netizens to share , True and reliable , And keep it updated in real time , At present, a total of 811 Dao Zhen （ Still growing ）. Support by company 、 department 、 Position combination queries the data in the warehouse , Convenient and efficient retrieval , Targeted response to interview .

Entering the homepage, we can see that it sorts the algorithm problems according to the frequency of investigation from high to low . You can see at a glance what you like in an interview .

Through the analysis of the question type of this website , We can clearly summarize the following three rules ：

• Data structure preferences ： Array & Hashtable > character string & Linked list > Binary tree & chart > other ;
• Difficulty tends to be simple 、 Medium class ;
• The investigation density is relatively concentrated ;

The following explains these three points in detail ：

1. Array & Hashtable > character string & Linked list > Binary tree & chart > other ;

I made a simple statistics , This website is from 2020 year 1 A total of 811 Algorithm questions , Where array & Hash table accounts for 226 Avenue （25.7%）, String takes up 95 Avenue （10.8%）, The linked list accounts for 102 Avenue （11.6%）, Binary tree 、 Figure occupation 120 Avenue （13.6%）, Stack & The queue occupies 84 Avenue （9.5）, There are also some miscellaneous math and design problems 76 Avenue （8.6%）.

In this way, the Internet has been covered 80% Question type of . The question types I listed above are actually the most basic data structures that you learned when you were an undergraduate . There is not even a sorting algorithm 、BFS/DFS, Let alone greedy algorithm dynamic programming .

therefore , Compared with going there 20% Kill all brain cells , First of all, we should 80% The score of is firmly in hand . Wait until the basic score is guaranteed , Then we can go to feel at ease to make the icing on the cake .

2. The topic tends to be simple 、 Medium class ;

Also speak with data , Difficult questions only occupy 103 Avenue （11.7%） What does this mean ？

Dachang algorithm questions do not require high difficulty . All in all 1 About hours , Algorithm examination can only account for 30min About time . For difficult topics , Most candidates can't guarantee to give a correct solution in such a short time . Therefore, the relatively simple and direct question type has become the only choice for examination .

3. The investigation density is relatively concentrated ;

You can observe the recent inspection time of these high-frequency topics , Is it yesterday , The day before yesterday , Even basically all the questions have been tested repeatedly within a week . What does this mean ？ It means that the question bank of the interviewer is highly overlapped and the update frequency is low ！

A little brother of an interviewer who had a good relationship told me , He only passed the interview this year 3 Problem , Over and over again .“ It's convenient ”, He said lazily ,“ I'm so busy at work , I don't have much time to study so many question types . It usually takes a little time to figure out a few classic questions in advance , Change later when you are free .” It looks like this brother's “ Available ” It's delayed .

So , The life cycle of high-frequency questions is permanent , Is enduring , It is used repeatedly . Maybe another company will take the exam if this company gets the exam . Pondering over high-frequency questions is a very time-saving and labor-saving trick .

Said so much , Let's summarize it for you . Give priority to High frequency 、 Moderately difficult 、 More basic data structure question type .
Promise me , In this step, make sure to do well , You don't need to use your brain completely , It only takes a little effort to successfully overcome 70% Interview Algorithm .

Yes, of course , It's not enough . You who chose to support sister Yien , IQ is definitely better than ordinary people （ Manual formation ） remainder 30% We have to work hard to win it .

Now let's talk about the specific methods of brushing questions ？

The sea of questions is vast , Ten million question banks . How can I choose ？

Are you still confused about this problem ？ Don't worry , I brushed it all at this time 2 It's a question . Just look at three ：

• The finger of the sword Offer（ It must be accompanied by books ！）
  • Question bank address ：https://www.nowcoder.com/exam/oj/ta?tpId=13
  • 《 The finger of the sword Offer》 Address ：https://blog.csdn.net/monk1992/article/details/96438796
• LeetCode
  • Question bank address ：https://leetcode-cn.com/problemset/all/
• The real problem of the company
  • Company real problems sorted out by Niuke ：https://www.nowcoder.com/exam/company
  • In fact, I prefer to turn over some partners' questions in the discussion area ：https://www.nowcoder.com/discuss

The location of resources is also posted on it , You can directly search on Niuke website . In fact, I highly recommend that you directly use Niuke's question bank to brush questions , its IDE And interface restore the real interview environment to a great extent . Let's get familiar with it in advance “ The atmosphere of the examination room ” It is very helpful for the feel and mentality of the interview that day ～

The main output of my question brushing is Jianzhi Offer,LeetCode And the company's real questions are all free to play auxiliary . For the finger of the sword , I brush every question carefully, at least 3 All over , Carefully summarize after brushing 、 The lines . and LeetCode It mainly aims at making special breakthroughs in some weak question types , Didn't brush systematically .

At the same time, one week before preparing for an interview with a company , I will brush a copy of the company's algorithm real problem every day three or four days in advance , It is mainly used to take chances , What if these two interviewers are copying the question bank with each other ？

Now let's start talking about the most important point ,

How to brush the question , Can we really master this problem ？

We will explain it one by one in three stages , Namely ： Before brushing the question 、 When and after brushing questions ;

One 、 Prepare before brushing questions

Everyone must abandon a wrong idea , Brushing questions is definitely not for meaningless numbers and rankings . But to master a type of question 、 To consolidate the data structure . So compared to saying “ I brushed it again yesterday 10 Problem ”, I prefer to say “ Yesterday, I got the two-way linked list through .”

So when you brush questions, you must brush classic questions , We should draw inferences from one instance , Brush again and again , Abstract a common solution from a class of problems , To be able to respond to all changes with constancy , This is the essence of question brushing , Not just to brush questions .

With such a general direction , What we should do in the first step is very clear ： Classify and prioritize . It's good to classify according to the category of data structure , More ready-made classification can be directly according to 《 The finger of the sword Offer》 The chapters of this book ; The priority can be sorted according to the question type rules summarized above . I made a classification, probably ：

Hashtables and strings > Linked list > Binary tree and graph > Binary search and binary sort tree > Stack 、 queue 、 Pile up > other （ Mainly some mathematical problems ）> recursive 、 Backtracking and partition & Greedy Algorithm > Search for > Complex data structures > Dynamic programming

Two 、 Think when you brush questions

I have my own set of boxing , Now let's explain the dismantling action .

First of all, you must brush the questions pertinently , After reading the topic, please ask a soul three questions ：

1. What kind of question does this question belong to ？
2. What is the solution of this kind of question ？
3. Is there a template that can be set ？

After clarification , Just take out the template and cover it directly . The template here is the template for you to summarize this kind of question type after brushing the question , The specific summary method will be mentioned in the following section .

If a set hits , So congratulations , The template you summarized is generic and appropriate ; If you find that there is no template or problem-solving ideas, don't worry . At this point, it's like facing a new algorithm problem and trying to solve it with the algorithm knowledge you have learned .

Don't be too strict about the speed of solving problems when writing problems for the first time , You can give it to yourself 2～3 Hours to slowly explore the best way .

Time goes by , Whether you solve it or not , It's time to end this problem . The next thing you need to do is to see others' solutions and ideas . If you are writing questions on Niuke , You can directly click on the... In the upper left corner “ Answer key ”.

Generally, the best solution is the one who likes most , We're going to see what's good about other people's code , What's the difference with your own code . Mainly from the idea of solving problems 、 Time 、 Space complexity and other angles , And whether we can get some general inspiration from others' code .

Standing on the shoulders of giants , To see further .

After absorbing the essence of others , You need to brush the question again . This time, I brush the questions to confirm whether you really understand and master the best solution . This time is successful AC after , Your question is really “ brush ” End .

3、 ... and 、 Summarize after brushing questions

This step should be the most critical .

I have been adhering to it since I was a student Summary is the core method to turn knowledge into your own use . Even at the working stage , Still very useful .

Summary of interview Algorithm , I'll go and classify by question type , And for each type of question, I will summarize a set of pseudo code .
Every time I finish a question , I will sum up the ideas summarized by this question under this type of question ：

1. First, let's take a look at the consistency between this idea and pseudo code . If it cannot be applied directly , Yes 2 One reason .1 Pseudo code is not general enough ,2 It really belongs to a new question type ;
2. If it is 1, We need to further improve our pseudo code according to the summarized ideas ;
3. If it is 2 Words , It can be used as a derivative of this type of question , Remember alone . The memory here is not to recite questions , But to summarize another set of templates .

Some people may wonder how to write pseudo code , Here I teach you a general method ： Since it is not easy to abstract solution , We abstract the topic . Simplify the topic into a knowledge point , Then write an algorithm for this knowledge point .

Let me give you an example ：
“ Sliding window algorithm ” Class question type can be used to solve the minimum covering substring 、 Find all alphabetic words in the string 、 The longest substring problem without repeated characters .

I summarized a set of abstract ideas of sliding window algorithm , You can respond to changes with constancy ：

int left = 0, right = 0;

while (right < s.size()) {
    
    window.add(s[right]);
    right++;
    
    while (valid) {
    
        window.remove(s[left]);
        left++;
    }
}

The summary idea is written in detail in this document , Come in and learn quickly ：https://blog.csdn.net/qq_29966203/article/details/100070662

Sometimes I find that I can't summarize a template for a certain type of problem , Or always stuck in a class problem . I will devote some time to tackling . At this time, I will come up with the sea tactics ：LeetCode. because LeetCode There are too many questions , have everything that one expects to find . I will be in LeetCode Search this kind of question directly on a big operation , I believe that after a lot of tempering , Brain and hand memory will also give you a set of your own methodology .

See here , If you are still a little confused about the summary , I don't know how to start .

Cough , The dry goods are coming now ？

After all, sister chongfan ruyien is devoted to landing for you , Fuck broken heart .

In the process of summarizing , according to 《 The finger of the sword Offer》 Our catalogue is divided into twelve categories , In each category, I start from the sword finger and LeetCode Selected several extremely representative question types （ It can also be said to be a knowledge point type question ） And write the corresponding solution , Each question is marked with LeetCode/ The corresponding question number on the sword finger , It is convenient for you to inquire and practice by yourself .

Family , Just rush and it's over .

Chapter one Linked list ：https://blog.csdn.net/qq_29966203/article/details/100496138

• leetcode
  • example 1： Reverse order of linked list （206）
  • example 1b： Reverse order of linked list 2（92）
  • example 2： Find the intersection point by linked list （160）
  • example 3： Look for links in a list （141）
  • example 4： List Division （86）
  • example 5： Replication of complex linked list （138）
  • example 6：2 Merge of sorted linked lists （21）
  • example 7：K Merge of sorted linked lists （23）
• The finger of the sword offer
  • example 1： Print linked list from end to end （3）
  • example 2： Last in the list k Nodes （14）
  • example 3： Reverse a linked list （15）
  • example 4： Merge two sort lists （16）
  • example 5： Replication of complex linked list （25）
  • example 6： The first common node of two linked lists （35）
  • example 7： Children's games （ The last remaining number in the circle ）（45）
  • example 8： Link entry node in the linked list （54）
  • example 9： Delete duplicate nodes in the list （55）

Chapter two Stack 、 queue 、 Pile up ：https://blog.csdn.net/qq_29966203/article/details/100550574

• leetcode
  • example 1 Use queues to implement stacks （225）
  • example 2 Queues are implemented using a stack （232）
  • example 3 contain min Function of the stack （155）
  • example 4 Legal stack queue （946）
  • example 5 A simple calculator （224）
  • example 6 No K Large number （215）
  • example 7 Look for the median （295）
• The finger of the sword offer
  • example 1： Queues are implemented with two stacks （5）
  • example 2： contain min Function of the stack （20）
  • example 3： Pressure into the stack 、 Pop-up sequence （21）

The third chapter Greedy Algorithm ：https://blog.csdn.net/qq_29966203/article/details/94612334

• leetcode
  • example 1： Distribute cookies （455）
  • example 2： Swing sort （376）
  • example 3： remove K A digital （402）
  • example 4a： Jumping game （55）
  • example 4b： Jumping game 2（45）
  • example 5： Shoot balloons （452）
  • example 6： Optimal refueling method
• The finger of the sword offer
  • example 1： Cut the rope （66）
• 2019 The school recruit
  • example 1： Place street lights （3）

Chapter four recursive 、 Backtracking and partition ：https://blog.csdn.net/qq_29966203/article/details/93963493

• The finger of the sword Offer
  • example 1：Fibonacci The sequence
  • example 2： Skip step
  • example 3： Abnormal step
  • example 4： Rectangular overlay
  • example 5： Path in matrix （64）
  • example 6： The range of motion of the robot （65）
  • example 7： Make the array the smallest number （32）
  • example 8： Arrangement of strings （27）
  • example 9： Reverse pairs in arrays
• leetcode
  • example 1： Count the number of elements on the right less than the current one
  • example 1： Finding subsets 1（ The original array does not contain duplicate elements ）
  • example 2： Finding subsets 2（ The original array contains duplicate elements ）
  • example 3： Bracket generation
  • example 4：N Queen
  • example 5： The matchstick is square （473）

The fifth chapter Binary tree and graph ：https://blog.csdn.net/qq_29966203/article/details/96424693

• leetcode
  • example 1： Sum of paths （113）
  • example 2： Recent public ancestors （236）
  • example 3： Binary tree turn chain list （114）
  • example 4： Look at the binary tree from the side （199）
  • example 5： course arrangement （ Directed graph judgment ring ）（207）
• The finger of the sword offer
  • example 1： Reconstruction of binary tree （4）
  • example 2： The substructure of a tree （17）
  • example 3： Image of binary tree （18）
  • example 4： Print binary tree from top to bottom （22）
  • example 5： The path of a value in a binary tree （24）
  • example 6： Binary search tree and double linked list （26）
  • example 7： The depth of the binary tree （37）
  • example 8： Balanced binary trees （38）
  • example 9： The next node of a binary tree （56）
  • example 10： Symmetric binary tree （57）
  • example 11： Print binary trees in zigzag order （58）
  • example 12： Print a binary tree into multiple lines （59）
  • example 13： Serialize binary tree （60）

Chapter six Binary search and binary sort tree ：https://blog.csdn.net/qq_29966203/article/details/100169820

• leetCode
  • example 1： Insertion position （35）
  • example 2： Interval search （34）
  • example 3： Rotate the array to find （33）
  • example 4： Binary search tree encoding and decoding （449）
  • example 5： Reverse order number （315）
• The finger of the sword offer
  • example 1： Minimum number of rotation array （6）
  • example 2： The post order traversal sequence of binary search tree （23）
  • example 3： The number of times a number appears in a sort array （36）
  • example 4： Binary search tree No k Nodes （61）
• 2019 The school recruit
  • example 1： Have bumper harvest （11）

Chapter vii. Hashtables and strings ：https://blog.csdn.net/qq_29966203/article/details/100029889

• leetcode
  • example 1： Longest palindrome （409）
  • example 2： Word pattern （290）
  • example 3： Group words with the same character （49）
  • example 4： Longest substring without repeating characters （3）
  • example 5： Repetitive DNA Sequence （187）
  • example 6： Minimum window string （76）
• The finger of the sword offer
  • example 1： Replace blank space （2）
  • example 2： A number that appears more than half the times in an array （28）
  • example 3： The smallest K Number （29）
  • example 4： The first character that appears only once （34）
  • example 5： A number that appears only once in an array （39）
  • example 6： Left rotation string （42）
  • example 7： Reverse the sequence of words （43）
  • example 8： Convert a string to an integer （48）
  • example 9： Repeated numbers in an array （49）
  • example 10： The first character in a character stream that does not repeat （53）

Chapter viii. Search for ：https://blog.csdn.net/qq_29966203/article/details/100088992

• leetcode
  • example 1： Number of Islands （200）
  • example 2： Word ladder （127）
  • example 3： Word ladder 2（126）

Chapter nine Dynamic programming ：https://blog.csdn.net/qq_29966203/article/details/95489974

• leetcode
  • example 1： climb stairs （70）
  • example 2： raid homes and plunder houses （198）
  • example 3： The largest field and （53）
  • example 4： give change （322）
  • example 5： triangle （120）
  • example 6： Longest ascending subsequence （300）
  • example 7： Minimum path sum （64）
  • example 8： Dungeon game （174）
• The finger of the sword offer
  • example 1： The maximum sum of successive subarrays （30）
  • example 2： Maximum sliding window （63）
• 2019 The real problem of school recruitment
  • example 1： Niuniu looking for a job （1）
  • example 2： Niuniu's backpack problem （8）

Chapter ten Complex data structures

Chapter 11 other ：https://blog.csdn.net/qq_29966203/article/details/100658666

• leetcode
  • example 1： Invert binary （190）【 An operation 】
• The finger of the sword offer
  • example 1： Two dimensional array search （1）【 Array 、 intelligence 】
  • example 2： Binary 1 The number of （11）【 An operation 】
  • example 3： Integer power of value （12）【 Code integrity 】
  • example 4： Adjust the array order so that the odd Numbers precede the even Numbers （13）【 Code integrity 】
  • example 5： Print matrix clockwise （19）【 Draw pictures to visualize the abstract 】
  • example 6： In integers 1 Number of occurrences （31）
  • example 7： Ugly number （33）
  • example 8： And for S Continuous positive sequence of （40）
  • example 9： And for S Of 2 A digital （41）
  • example 10： Playing card shunzi （44）
  • example 11：1+2+3+…+n（46）【 Conditions and && Short circuit principle 】
  • example 12： Do not add, subtract, multiply or divide （47）【 An operation 】
  • example 13： Building a product array （50）【 Array 】
  • example 14： Regular Expression Matching （51）【 simulation 】
  • example 15： String representing the value （52）【 simulation 】
  • example 16： Median data flow （62）
• 2019 The school recruit
  • example 1： By 3 to be divisible by （2）【 mathematics 】
  • example 2： Lost Cow （4）【 simulation 】
  • example 3： Number pair （5）【 mathematics 】
  • example 4： Overlapping matrix （6）【 mathematics 】
  • example 5： Niuniu's alarm clock （7）【 date 】
  • example 6： tetris （9）【 simulation 】
  • example 7： doze （10）【 simulation 】

Refined above about 120 About algorithm questions are classic .
I brushed every question at least 3～4 All over , The summarized template can let me have full confidence to deal with all interview questions . meanwhile , For the short time before the exam, how to cram fast and efficiently , I will further condense and then condense the above algorithms , Summarized a simplified version of algorithm summary and induction , common 50 No problem . Basically, I spend about half an hour thinking about the algorithm model before every interview , Keep a code hand （ brain ？） sense .
I won't list the catalogue , Let's click inside . Each question already belongs to a core compulsory knowledge point .
Algorithm summary and induction
https://blog.csdn.net/qq_29966203/article/details/102745404

And because the topic summarized is too important , So I once again printed out these algorithm problems , Recite once a day , Happy yangleduo ～


I will also after every interview , For this interview, we need to re consolidate and annotate our own note taking system . So the printed notes are dense , Some places can't remember, so they stick post it notes ……

One more thing to say , Summing up is really a very good study habit , He can really turn knowledge into his own use , Be your own . After reading Yi en's information , Be sure to summarize your own information , Summarize your own template . I believe that after having a Wulin secret script of my own , No matter what algorithm problem you face , Can be evil spirit crazy smile “ o , Is this ？”

Okay , The method of writing questions is almost the same , Finally, I've come up with some from various interviews

Hand tear code super easy to use small tips～

1. First of all, whether there is an interview or not , Make sure you brush it all day 2～3 topic , Don't take long , Give yourself at most half an hour for each question . Then look at the solution , summary . Don't care too much whether you brush it or not , The key is to cultivate a touch .
2. Go over the core algorithm questions and ideas you summarized before the interview , Make sure the template routine is kept in mind and applied freely
3. When the interviewer asks you to tear the code by hand , it is to be noted that ：
4. Be sure not to rush to write ！ Even if you hear the question halfway, you will find that you have brushed it , Don't rush up in surprise public void main！ modesty 、 Can you keep a low profile ？ This is the time , You have to show your lifelong acting skills , Pretend to frown and think for a while , Then ask the interviewer if his idea is right , Write after reaching a consensus ;
5. Pay attention to neat code format and writing habits when typing code . Every line of code you type during the interview is exposed to the interviewer , Whether the figure is good or bad can be seen at a glance ;
6. For some strange or thoughtless questions, you must not be anxious or sit still , Take the initiative to talk about your ideas or discuss with the interviewer . Interviewers usually give some hints , These tips are often critical , And it's a turning point . After all, there is hope when you knock it out , It's generally cool if you can't knock it out .

At the end

The hand tearing code of interview , Compared with the algorithm problems you usually participate in various competitions , The biggest difference is that it is too simple . Many people think it difficult , In fact, the problem is not difficult , It's a kind of tension created in the one-on-one atmosphere of interview 、 The suffocating atmosphere will greatly reduce your good code ability , Many questions . So we must maintain a good attitude , At the same time, you should brush more questions , Try to replace your brain with inertia . Prevent some difficult accidents .

Finally, if this article is helpful to you , Remember to click a like or collection ～ Sister Yien wishes all the little cute people who like it can successfully go ashore ～