Projections in ClickHouse

Mark Needham

Discover how to dramatically improve query performance in ClickHouse using projections - a powerful feature that creates query-optimized copies of your data.

This video demonstrates projections using a real-world example with 30 million UK property price records. Watch as we reduce query time from 53 milliseconds to just 4 milliseconds by creating a projection for calculating average house prices by county.

What You'll Learn:

• What projections are and how they work as hidden tables that update atomically
• Three main use cases: reordered datasets, subset of columns filtering, and pre-computed aggregations
• Step-by-step process to create and materialize projections
• How to verify projections are being used with EXPLAIN queries
• Understanding the storage trade-offs and space requirements