ClickHouse matches PG for single-row UPDATEs and 4000 x faster for bulk UPDATEs

Summary

TL;DR· On identical hardware and data, ClickHouse matches PostgreSQL for single-row UPDATEs and is up to 4,000× faster in our tests for bulk UPDATEs. · Why it matters: Bulk updates are common in OLTP workloads, and ClickHouse’s columnar design + parallelism make them far faster. · Caveat: PostgreSQL is fully transactional by default; ClickHouse isn’t. Results compare each engine’s native execution model, not identical transaction guarantees. PostgreSQL is the most popular open-source OLTP database in the world. It’s a fair assumption that, when a developer thinks about UPDATE performance, they probably think of the baseline that PostgreSQL sets. In contrast, ClickHouse is the most popular open-source OLAP database in the world. While ClickHouse sets the expectation of OLAP performance, its OLTP potential is often underrated. In July 2025, ClickHouse v25.7 brought high-performance SQL-standard UPDATEs to ClickHouse. Naturally, we want to know how ClickHouse stacks up against the best. Are UPDATEs in ClickHouse competitive with UPDATEs in Postgres? In our previous post, we benchmarked the new high-performance SQL-standard UPDATEs against other ClickHouse update methods. Here, out of curiosity, we’re running the same benchmarks against PostgreSQL on identical hardware and data. We focus on two OLTP staples: single-row updates and bulk changes on a classic orders dataset. These tests reveal how much update speed depends on finding rows efficiently, a strength ClickHouse inherits from its analytical roots. Before diving into the numbers, we want to acknowledge something: a PostgreSQL vs. ClickHouse comparison isn’t apples-to-apples. Both can run UPDATEs, but their execution models differ in ways that matter when interpreting results. TL;DR: PostgreSQL wraps every statement in a fully transactional context; ClickHouse doesn’t. That means these results aren’t a perfect measure of transaction performance. However, they’re still an interesting and relevant look at how each s...