ClickHouse Interview Questions

ClickHouse is built for OLAP: analytical queries over huge volumes of data (billions of rows), columnar storage, and high scan throughput. Use it for dashboards, logs, events, time-series analytics. PostgreSQL is row-based and optimized for OLTP: transactions, updates, point lookups. Use PostgreSQL when you need ACID, frequent updates, and complex JOINs on smaller datasets. Different tools for different jobs.

MergeTree is ClickHouse’s primary table engine for mutable data. Inserts create data “parts”; a background merge process combines parts over time. You specify ORDER BY (and optionally PARTITION BY) for how data is sorted and split. MergeTree gives you fast range scans and supports TTL, deduplication, and various merge strategies (e.g. ReplacingMergeTree, SummingMergeTree). Most production tables use a MergeTree-family engine.