How to PASS SQL Interviews? Data Science Tips⚡
ฝัง
- เผยแพร่เมื่อ 30 มิ.ย. 2024
- Become an SQL pro with sql.datainterview.com/ 🚀
====== ✅ Details ======
SQL interviews in data science are tricky. Even seasoned pros fail the SQL rounds at times.
That's why in this video, I share:
👉 Tips that make or break your success in SQL interviews
👉 Daetama - an SQL practice pad with 100 interview questions
👉 A walkthrough of a Google SQL interview question
⭐ Daetama is a new DataInterview product that will allow you to practice code on your web browser. Get 100 FAANG-style SQL questions solved by data scientists at Google and Meta. Visit daetama.io/
🚀 Looking for data science interview prep? Check out: datainterview.com/
👍 Make sure to subscribe, like, and share!
====== ⏱️ Timestamps ======
00:00 - How to PASS SQL Interviews?
00:35 - What do interviewers look for?
00:58 - Google SQL question
01:33 - Bad Response
02:12 - Good Response
04:49 - Great Response
====== 📚 Other Useful Contents ======
1. How to Ace Product Metric Questions 👉 bit.ly/3xeCgOl
2. Cracking Data Science Business Cases 👉 bit.ly/3trCHDP
3. Crack the Amazon Data Scientist Interview 👉 bit.ly/3MyC6XJ
====== Connect ======
📗 LinkedIn - / danleedata
📘 Medium - / datainterview
Hope you enjoy today's video on SQL! Got questions? Drop one below!👇
Please keep up the great work you're doing! Super helpful video!
2 things I didn't understand:
1. Why didn't you do a left join? even if the interviewer doesn't expect to see websites without clicks, it's still a better answer that shows that you think of all edge cases.
2. Why do you have to use a WITH or an inner query? I think a better solution would be to ask the interviewer if it make sense to assume that the websites table is rather small (not more than 100k probably), therefore you can easily join it with the events table without losing computation, and then do the join, filter and group by in one simple query.
Thank you for the videos! I believe it would be beneficial to know how the solution would change (if at all) depending on the answers given by the interviewer in response to the interviewee's follow-up questions.
I'm excited about tomorrow's launch of Daetama at 9am !
And, Daetama has launched :)
Thank you for keeping updated with such informative content and great work with Daetama launch. I want to practise more questions as I started following your videos lately. Seems like this coupon code got expired, I humbly request you to please enable it.. 🙏🏼 Thanks in advance!!
There's another code that's active right now. It's DATAINTERVIEW10
Hi Dan, I have a habit of checking results output of each step before reaching the final completed code.
For example, I would check the result of my CTE, or subquery first, before joining. Is that allowed in the interview?(if it's not on the phone screen) Thanks
That depends on the coding environment. For places like Google and meta, you write sql on a non-executable environment because they want to see whether you have the logic internalized. For others, if they let you execute, heck why not :)
Yup, not being able to execute a query through out the process would drive me crazy.... but I guess we need to be ready for that.
I have a question. What is a standard (if exists) amount of time that is given to the interviewee per SQL question to solve? And if it varies a lot, what is a "safe" (subjective) time limit that I should strive to constrain to for practice purposes?
Hey Justin, generally for technical screen about 20 minutes of the 45 minute round is reserved for SQL, and about two to three SQL problems are asked. I would say that on average it should take you about 7 to 10 minutes to solve each problem. When I self-practice, I usually aim for 5 minutes at tops.
@@DataInterview Thanks for the quick response! I am pretty skilled with SQL, and just wanted to have an idea of time limits since I haven't had a SQL interview before. 5 minutes seems like a good standard for practice, I will go with that. Thanks again.
Cool!