Реальное собеседование на Data Engineer в Wildberries на 400 000 рублей | Собес в IT

Telegram channel: https://t.me/+_XJLCGNES0s4MjU6 Sign up for training: https://t.me/m/fHpdGrTSNGNi 0:00 — Introduction 0:36 — What's the difference between union and union all? 1:24 — If there are duplicates in a section, for example, above or below union, will they be retained or removed? 2:29 — Problem 1 4:26 — What other window functions have you worked with, and what problems have you solved with them? For example, using raw numbers. 7:58 — Let's say Rank and Dense Rank are different? 9:16 — I have a table with two columns: first name and last name. I need to clear the table of duplicates. 13:27 — Tell me about Clickhouse; how does its primary key differ from the primary key of classic OLTP systems? 15:26 — Are duplicates possible in a primary key in Clickhouse? 16:23 — Problem 2 18:30 — What if we solve the problem not on an OLTP system, but on a columnar system, for example, on Clickhouse? 20:43 — What happens under the hood with a large number of small inserts in Clickhouse 23:37 — What happens if we have too many small parts? 25:00 — Does Clickhouse have any mechanisms to optimize the handling of these parts? 27:06 — You say Clickhouse will take a long time to read small parts. What does it do with them? 29:00 — How is a broadcast join different from a shuffle join? 31:51 — How are Clickhouse's mat views different from classic relational mat views? 36:01 — Tell us about the copy-on-write and merge-on-read data management strategies. What is it and in what cases is it better to use which of them?