Unlock this content
Enter your email to unlock this content for free
Intro to Engines
TL;DR
Table engines define ClickHouse's data architecture: how data is stored on disk, indexed, merged, and queried. They control partition layout, merge behavior, replication, and query performance. Start with MergeTree for most production use cases.
The table engine defines the data architecture: storage (partition layout, columns, indexes), lifecycle (how parts are created and merged), and query behavior (how data is read and aggregated).
How Engines Work
Different engines change table behavior completely. The same schema with different engines behaves differently.
-- Memory engine: data in RAM only, lost on restart
CREATE TABLE temp_table (
id UInt64,
name String
) ENGINE = Memory;