So I’ve been looking into SQL optimization and I’m curious what people actually dealing with production databases think.
I’m not trying to make some magical database engine that beats everything at everything lol.
What I’m more interested in is whether there’s a specific type of SQL operation that current engines still struggle with, even after all the usual optimizations.
Like if companies are spending ridiculous amounts of compute on certain joins, aggregations, sorts, or whatever else, why not investigate whether there’s a fundamentally better way to handle just one of those operations?
I’m talking about actual algorithmic improvements too, not just adding indexes or tweaking query plans.
Couple things I’m curious about:
1. What SQL operations actually eat up the most CPU or money in your workloads?
2. Are there specific operations that PostgreSQL, DuckDB, ClickHouse, Snowflake, etc. still struggle with?
3.…