Where does parallelism appear in Airflow?
Data Engineer
Tell me, are you familiar with the Domain Driven Design (DDD) methodology? If yes, share examples of its application in your projects.
Analysis of fitness club attendance You work as an analyst in a chain of fitness clubs. You have information about user visits and the memberships they purchase. It is necessary to analyze the effectiveness of membership usage. Calculate for each type of membership: • the total number of users who used this type of membership. Consider only unique user_id; • the total number of visits for this membership. Consider all visits of users with this membership; • the share of users of this membership in percentage of the total number of all users (rounded to one decimal place). To calculate the share, use the ratio of the number of users with this membership to the total number of all unique users. Each user can have only one membership. Sort the result by membership type in alphabetical order. Input format Memberships table: • membership_id (int) — unique membership identifier • user_id (int) — unique user identifier • membership_type (text) — type of membership Visits table: • visit_id (int) — unique visit identifier • user_id (int) — user identifier • visit_date (timestamp) — date and time of visit Data does not contain missing or incorrect values. Output format The query should return a table with fields in this order: • membership_type (text) — type of membership • users_count (int) — number of unique users with this type of membership • total_visits (int) — total number of visits of users with this membership • user_share (numeric) — share of users of this membership in percentage of the total number (rounded to 1 decimal place) The result is sorted by membership type in alphabetical order.
How did you work with Impala?
SELECT * FROM table1 AS t1 LEFT JOIN table2 AS t2 ON t1.date_start > t2.date_start
What was your role in a team of 4 people, and how do you see your career in 3-4 years, what goals do you set?
What are indexes in databases, why are they needed, and what is their internal structure?
What are your motivations and development plans for the next few years? Where do you want to grow?
SELECT * FROM table1 AS t1 LEFT JOIN table2 AS t2 ON t1.date_start > t2.date_start d = { True: 42, 1: 2, 1.0: 100500 } print(d)
CREATE TABLE orders ( driver_id varchar, city varchar, order_id varchar ); -- Get the top 10 drivers by number of orders in each city.
Tell us about your experience with Greenplum, DBT, and Data Vault. Why did you switch from a snowflake model to Data Vault?
Tell us about your experience, the tasks you solved, and the stack.
If an SQL query takes too long, what will you do?
Where is data compression more effective — in columnar or row storage, and why?
Tell us about your work experience: tasks, technology stack.
Have you worked with Oracle and PL/SQL?
What is shuffle in Spark and why is it needed?
What is the difference between iterating over a string and iterating over a tuple in Python? What are the basic types and how does this affect performance?
Why did you decide to leave your current project? What is unsatisfactory there?
Threading, multiprocessing, asyncio: what is the difference and when is it better to use each?