Amazon SQL Mock Interview Question: Conversation Distribution

Today I'm doing a mock SQL interview with Jitesh, a data engineering coach on Interview Query, with 10 years of experience in data engineering and analytics at companies like Databricks and Amazon.
Here's the question we're tackling: www.interviewq...
Want to be featured in the next mock interview video? Apply here: airtable.com/s...
👉 Subscribe to my data science channel: bit.ly/2xYkyUM
Use the code "datasciencejay" and get 10% off data science interview prep 🔥 : www.interviewq...
❓ Check out our new SQL interview course: www.interviewq...
🔑 Get professional coaching from Jitesh here: www.interviewq...
🐦 Follow us on Twitter: / interview_query
Chapters:
1:15 - What's the distribution look like for all conversations?
3:00 - Write a query to get the distribution
4:34 - Calculating the frequency
More from Jay:
Read my personal blog: datastream.sub...
Follow me on LinkedIn: / jay-feng-ab66b049
Find me on Twitter: / datasciencejay

[@_sds_1499]

Not convinced with the solution. Eg: say we have 2 records : 1,2 and 1,3 for user 1 and 2 columns respectively. User 1 will have 2 messages. But user 2 and 3 will also have a message each. In the video approach, only 2 messages of user 1 will be counted. In fact, this approach will succeed if the records contain double counting

[@poojasikdar3583]

I'm just wondering if number of conversations here represents distinct users a person interacted with. I might only be interested in conversations a user had per day. Other than that I think the solution was to the point.

[@observer698]

I wonder if the definition of a number of conversations is the total number of msg_count or number of people someone interacts with in a day.

[@JTChi]

Jay I am a member of Interview Query. Where exactly can I find these interview questions?

[@iqjayfeng]

www.interviewquery.com/questions/conversations-distribution

[@JTChi]

@@iqjayfeng Thank you

[@everettehungerford2858]

So, I'm learning SQL and just want to ask a question. How can you count all distinct instances of user 2 as being representative of all conversations. Why not have the number of conversations be based off a distinct count of concatenated user columns?

[@poojasikdar3583]

Or just take the id?

[@TheElementFive]

Wow, the interviewee didn't pause to contemplate the question at all before jumping into the code..

[@businessmotivation2]

which platform is it? for practising these questions

[@iqjayfeng]

www.interviewquery.com/

[@LS-tj2ho]

He didn’t answer the question.

[@andrewchen7342]

So hard to understand what the interviewee was murmuring throughout the whole interview...

[@scw806a]

part of the difficult interview is when there is a heavy accent guy where you might need to ask "I'm sorry can you repeat your question again, then the guy got annoy and just disqualify you for the interview" .