Расскажи про текущее место работы: основной продукт, твой функционал.
Data Engineer
Which expression in place of the missing [...] will automatically create an index? There is a horizontal code scroll on the mobile platform create table some_table( col_name [...] ); unique references other_table(col_name) not null serial integer check (col_name > 0)
How do you verify data correctness when translating a pipeline from SAS to DBT?
dict1 = { (1, 2), [3, 4, 5] : 0 } var = 1, 2
What was the last task you solved in production using Python?
How long does a solution using the built-in sorted function take? Can a faster solution be devised?
Tell me what you know about columnar and row-based data storage. When and which one should be chosen and why?
Why are you leaving your current job?
Describe the methods of SQL query optimization you have applied. How did you reduce the execution time from 12-15 seconds?
You found that there are commits containing critically important confidential data in the main repository's history. These data need to be completely removed from the entire repository history. Evaluate how correct and safe it would be to use the following strategy: create a new commit that removes the confidential data from the current version of files, and push it to main. - Correct but not optimal. Better to use git revert to undo commits - Conditionally correct. This is a temporary solution until a more radical means of data removal is found - Incorrect and unsafe. Data will be removed from the current version but will remain accessible in the repository history - Incorrect. Such a commit may lead to new conflicts when merging with other branches - Correct and safe. This method guarantees that data will be removed and will not appear in the repository anymore
What is CTE (Common Table Expression)?
Swap files — how do they affect performance? When else can swap files occur?
LEFT JOIN: a table with 10 records LEFT JOIN a table with 100 records. What is the minimum and maximum number of rows that can be obtained?
Compare the Data Vault approaches in MCHS and X5. What benefits did the new objects (bridge, pit tables) provide?
What data quality checks were performed in Data Vault?
What causes deadlocks and which mechanism in databases is responsible for data consistency during parallel queries?.
What is partitioning and sharding (distribution)?
We input metadata, and it creates a flow based on this metadata — how will this help quickly migrate existing flows from old systems to the planned one?
You are working on a new feature in the dev branch. Suddenly, there is a need to urgently switch to the main branch to quickly fix a typo in the README.md file. You have some uncommitted changes in src/feature.js (unstaged) and styles/main.css (staged). You want to temporarily save all these changes to return to them later in the dev branch.
How does the granularity size affect the ClickHouse table?