Other DBMSs have similar typecasting to PostgreSQL, but they're all different. It's one of those things that never really standardised. Personally I like the br

Other DBMSs have similar typecasting to PostgreSQL, but they're all different. It's one of those things that never really standardised. Personally I like the breadth of different types available in PostgreSQL. First-class enums and ranges really help a lot of apps, and while table inheritance has issues it does have some really interesting uses. As for features to look in too, I'd look at the in-server functions (plpgsql, plv8, plperl etc) for extensibility. Custom and complex types are pretty awesome as well. Also the range of functionality in window functions can make some really interesting queries possible. Also the concurrency model does need some understanding, as it means a commit will fail rather than deadlock. Why should use use PostgreSQL over Hadoop? If you have less than ~20TB of data then it's much less complex to set up and maintain, and as long as you don't want to read the entire data set at once it's fast. If you have more but it's easily partitioned (eg multi-tenant) then Citus have a package that can help, or if you have column-oriented data Citus have a different extension to help. Postgres is also really well tested and has decades of production use behind it, so backups and restores are well known, though that does require manual scripting to copy the files.