SELF-ASSESSMENT
Review Questions
1
2
Why does a query on 3 out of 30 columns return only about one-tenth of the data? Why has the `event_type` column shrunk by a factor of 211?
How do I choose the ORDER BY clause for an event table? Why did the filter based only on player_id return 218 granules instead of just one?
3
How is PARTITION BY different from ORDER BY? Why is PARTITION BY player_id an error?
4
What happens with every INSERT? Where does the "too many parts" error come from, and how can it be avoided?
5
When does `ReplacingMergeTree` remove duplicates? How can I get the correct count before the merge?
6
Why is the MV created after loading empty? What happens to the amount in the MV when the insertion is repeated?
7
8
How does a mutation differ from a lightweight DELETE and TTL? Why are frequent UPDATES a bad sign?
Which data from your project will remain in PostgreSQL, and which will be moved to ClickHouse? Under what measurable condition?
Question 8 — Start of the practice session. Sample answers — in the lecture notes, Section 16.