I think the biggest improvement/effort thing you could look at for Postgresql is the cost estimates. It really does a poor job at estimating row counts (over/u

I think the biggest improvement/effort thing you could look at for Postgresql is the cost estimates. It really does a poor job at estimating row counts (over/under estimating by multiple orders of magnitude quite often) when there are multiple joins, and the new extended statistics are for the most part useless for real scenarios that I have come across.