| Two features I'm excited about, aren't in this release, but are on the radar: 1. Incremental materialized view updates - enables efficient queries on denormalized views of normalized schema (because we don't need to rebuild a view for each update) ... Anybody know some good ways to do this right now? 2. Integration of openCypher / GQL stuff into SQL. Graphs and Sets make a lot of sense together and Apache AGE demonstrates this by letting you do stuff like join across multiple graphs... Further random ideas: It's fun to imagine how to integrate some of the best of other query languages like Datalog into PostgreSQL. That could be extremely useful for logic-based access control PL/Rust could be fun, too. I like the idea of using macros to generate lots of code, especially type definitions across multiple languages. One thing that isn't on the radar for PostgreSQL which I would love to see more rigorous debate on: declarative schema. The imperative migrations are super useful but I wish we didn't have to jump through so many hoops to have zero-downtime migrations. If SQL is declarative then the table definitions ought to have a declarative option as well, then we can just maintain the current desired state of the database schema in source and auto-migrate with CI/CD. I realize there are issues with possible incorrect diffs but, it would be worth it for simplicity. Just wishing for "Terraform for database schema" Thanks so much to everyone who works on PostgreSQL |
--> https://github.com/zombodb/plrust "A Rust procedural language handler for PostgreSQL"
and PL/Julia work in progress:
* GSOC2021 "Create procedural language extension for the Julia programming language"
* https://summerofcode.withgoogle.com/projects/#67439132530442...