I took two online written exams last week: iQIYI and Meituan, both of which were the background directions I chose, while I still have some fragmentary memories, I wrote my own summary.
iQIYI
The online written test of iQIYI is from 19.00-21.00 on 2.19, a total of two hours, 20 multiple-choice questions and 3 programming questions. Let’s talk about this process. The 20 multiple-choice questions are not difficult. They all ask some basic knowledge. The overall feeling is not difficult. Those who have brushed some questions should be able to answer them relatively easily. Let's talk about the 3 programming questions, I only have a 100% pass rate for the second question, and the remaining two questions are 40% and 30% respectively. After the online written test was completed, I went to the Niuke.com discussion area to see the answers to the programming questions (PS: The answers here refer to those made by people who took the written test like me). It is also written in Java. I looked at other people's solutions and my own. It is simply the gap between heaven and earth. I remembered that when I was studying software engineering, the teacher explained basic algorithm problems to me, and then gave me three kinds of answers, one was written by students, one was written by programmers, and the other was written by engineers. It felt like the algorithm I wrote was written by a student, and the person who gave the answer was like the answer written by an engineer. Because I usually use less algorithms, I don’t participate in algorithm competitions myself, so I haven’t practiced algorithms, and I still write some commonly used algorithms. The moment I saw other people’s answers, I felt that there was a big gap between the algorithms. And I believe that people with good algorithms will not be too poor in foundation. I ran the programs for the remaining two questions separately and found that I didn't understand the questions at all. Yes, that's right, I didn't even understand the title of the other two questions. In layman's terms, the question was judged wrong. But after thinking about it carefully, even if I checked all the questions correctly, at most two questions, the third question is only 30% pass rate, because the third question is dynamic programming, well, this algorithm used to be only I have understood the idea, but I have no experience in how to apply it in real life scenarios. Overall the written test was easy, but I did poorly.
Meituan
20, a total of 140 minutes, 20 minutes more than iQiyi, 20 logic test questions, 30 basic knowledge questions, and 2 programming questions. Meituan has significantly more questions than iQiyi, and the overall difficulty feels comparable. Meituan’s basic knowledge is more difficult, but the two programming questions are relatively simple. Let’s talk about the logic test questions first. There are a lot of math problems in them, which are probably at the level of a normal junior high school student. The estimated time is 20 minutes, about one question per minute. I did it for about 40 minutes in total. It's not that I'm bad at math, but it's just that I'm a little uncomfortable if I haven't done these math problems for a long time. If I come from junior high school, the score can be higher and the speed can be faster. The math problems in the logic test are not difficult. Look for the patterns. It feels like an IQ test question. Look for the patterns and look at them. If you can't find them, choose randomly. There are 30 basic knowledge questions. Meituan estimated the time to be 50 minutes. I finished it in about 40 minutes. After all, it is a multiple-choice question, and it does not involve a fast calculation. This basic knowledge multiple-choice question of Meituan is obviously more difficult than that of iQiyi. It asked several questions in C language. Although I know some, but for some questions, because I haven't used C language for a long time, and I haven't studied it in depth. However, I really don't know how to choose. But for those answers, I think the compiler knows the answer as soon as it runs, and can also infer the compilation principle by the way (PS: I personally think the test site is here). There are also operations on Linux, although I think those operations may be very common, but I have never used them, such as: soft link creation command (ln -s [source file or directory] [target file or directory]), which I don't know. The common Linux commands (pwd, ls, etc.) have been used and known, and a little deeper is the knowledge blind spot, and it is also one of the knowledge points that I need to reserve knowledge next. The passing rates of the two algorithm questions of Meituan are 30% and 40% respectively. The first question is because of the blind spot of knowledge. I don’t know what the gcd (greatest common divisor) is. At that time, it was quietly Baidu, maybe because of panic , and no results were found. If the first question gcd is known, my pass rate is 100%. The second question is even more pitiful. The pass rate has dropped from 40%, 30%, 20%, and 10%. It has been revised once and dropped once. The latter is really helpless. Finally, the 40% pass rate of the first version was submitted. went up. Because this question has involved data overflow, that is, the int representation can no longer satisfy this question, and long is needed. Java's int is 32 bits, which is enough to receive input data. The maximum input data is 10^9, but it is not enough to receive the output of the calculation result, so long must be used. But my problem is not here, because considering the data overflow, I use long all the time. The error is that the details are not considered enough, or it is a kind of thinking limitation. The solution, suddenly understood.
The difficulty of the two written tests is divided by the subjective feeling of answering the questions. Maybe for some big cows, the difficulty of the two sets of test questions is average. I will make up for the deficiencies reflected in these two online written tests in the next study, and strive to not ask me about my knowledge blind spots in the future interviews (not at all), and I need to show that I understand it.
Going farther and farther, I hope to look back one day, the past is light and cloudless!