Post #3141472

Optimising DuckDB performance on large EC2 instances "I have been recommending for some time that DuckDB is often a faster and simpler choice than Spark for most data engineering workloads. (...) However, there are some performance footguns with this approach: the default configuration can result in surprisingly poor performance, and a few simple changes can make workloads run multiple times faster." #duckdb https://www.robinlinacre.com/optimising_duckdb_performance_large_ec2_instances/