Systems

This section maps the directories under src/ to the engineering subsystems they implement. Each subsystem has its own page; this index is a roadmap.

If you are new to the codebase, read Architecture first, then drill into the systems below.

Storage

• Storage engine — the dispatcher (IStorage, StorageFactory) and the catalog of engines.
  • MergeTree — the columnar engine that powers analytics.
  • Replicated MergeTree — replication via Keeper / ZooKeeper.
  • Other engines — Distributed, Memory, Buffer, Kafka, S3, MaterializedView, Iceberg, etc.

Query path

• Query pipeline — the journey from SQL string to columnar response.
  • Parsers — src/Parsers/ (SQL → AST).
  • Analyzer — src/Analyzer/ (AST → QueryTree).
  • Planner — src/Planner/ (QueryTree → QueryPlan).
  • Interpreters — src/Interpreters/ (DDL, joins, aggregation, context).
  • Processors — src/Processors/ and src/QueryPipeline/ (the executor).

Computation

• Functions — src/Functions/ (the scalar function library).
• Aggregate functions — src/AggregateFunctions/.
• Data types and columns — src/DataTypes/, src/Columns/, src/Core/.

I/O and storage substrate

• IO and disks — src/IO/ and src/Disks/.
• Compression — src/Compression/.
• Formats — src/Formats/ and src/Processors/Formats/.

Coordination, replication, distribution

• Coordination — src/Coordination/ (the Keeper engine).
• Backups — src/Backups/.

Server-side

• Server protocols — src/Server/ (HTTP, native TCP, MySQL, Postgres, gRPC, ArrowFlight).
• Client library — src/Client/ (shared with clickhouse-client/-local).

Cross-cutting

• Access control — src/Access/.
• Dictionaries — src/Dictionaries/.
• Common utilities — src/Common/ (allocators, hash tables, logging, ZooKeeper client, …).

Other top-level directories

• src/Daemon/ — base daemon (signal handlers, config reload, log rotation).
• src/Loggers/ — log channel configuration.
• src/BridgeHelper/ — clients for clickhouse-odbc-bridge and clickhouse-library-bridge.
• src/TableFunctions/ — file(), s3(), url(), cluster(), iceberg(), delta(), etc.
• src/Databases/ — the database catalog (Atomic, Ordinary, Replicated, Lazy, MaterializedMySQL, MaterializedPostgreSQL).
• base/ — the project's compatibility/base layer (glibc-compat, Poco fork, pcg-random).

See also

• Apps — the binaries that wrap these systems.
• Features — cross-cutting capabilities expressed in user terms.
• Reference → data models — the data-type catalog.