[loic-sharma]

This is incorrect. It was their query engine that was hard to optimize, not the language. InfluxDB has been working on a new query engine based off Apache DataFusion to fix this.

If you squint, this query language is very similar to Polars, which is state-of-the-art for performance. I expect Pql could be as performant with sufficient investment.

The real problem is that creating a new query language is a ton of work. You need to create tooling, language servers, integrate with notebooks, etc… If you use SQL you get all of this for free.