Hey all. I've just pushed dataset 1.6.0, which pins the sqlalchemy dependency for this library to >= 1.3.2, < 2.0.0. This is meant as a hotfix to prevent people from getting broken installs by default.

Beyond that, ensuring that dataset will run with both sqlalchemy 1.3, 1.4 and 2.0 will be a bit of work - not horrible, but a solid afternoon of coding. It's forced me to realise that I haven't been running dataset in any of my projects for over two years, and this is also starting to show in some of the developer tooling the lib is using. So I'd like to send out a ping to the community and see if anyone is interested in taking the lead in adapting the library, perhaps even with a long-term view to becoming its new maintainer.