ClickHouse crash course. Conquer big data with ease

Build a proof-of-concept for ClickHouse by setting up a server and client with Docker-based configs, managing ports, and testing connections via Starbucks.

Introduce the merge-tree engines family and determine when to use them, highlighting their replication capabilities. Cover partition and order by options, including partition by expression and detaching partitions.

Benchmark merge-tree engines by simulating inserts and selects across configurable workers and batch sizes, comparing partition by, order by, and primary key configurations to reveal performance trade-offs.

Explore collapsing merge tree in ClickHouse, observe how inserts and updates affect data parts, and learn when final or canceling behavior impacts data collapse and query results.

Learn how ClickHouse dictionaries connect to external sources like databases and CSP files, offer fast lookups, and support flat dictionaries with unsigned 64 keys up to 500000 elements.

Explore sampling in ClickHouse by enabling it on tables, testing by key with order by, and benchmarking memory usage; findings show no clear performance gains, but memory consumption can decrease.

Explore arrays in clickhouse design by mapping names with array map, unfolding arrays with array join, and transforming to a normalized, one-to-many relational view without relying on slow joins.

Explore ClickHouse aggregates that work out of the box. Use materialized views and aggregate functions to compute sums and unique counts, speeding queries while keeping ingestion efficient.

Explore setting up a ClickHouse cluster with replication and sharding using docker compose and keeper, with macros to simplify distributed tables and data distribution.

Master conquering big data with clickhouse through a fast, agile course that progresses from basic to intermediate, sharing building blocks and code snippets while inviting community dialogue and feedback.