My database of choice is PostgreSQL. I often find myself wanting to do quick ad-hoc queries against a set of well-defined databases. Maybe I just want to know the internal id of a customer from their name, or check on the status of a job.<\/p>\n\n\n\n
If I’m doing real reporting and producing outputs I will usually spin up PyCharm. It’s got amazing db support, which is the same as you would get from JetBrain’s dedicated database product DataGrip. Great ergonomics, but Pycharm takes a while to get started and frankly the setup of all our dbs is onerous enough that I haven’t registered all of them. I also use pgAdmin<\/a>, but that’s also tedious to set up “just right”. And sometimes when you’re already in a terminal mindset, you don’t want to jump out to a GUI.<\/p>\n\n\n\n So I used to ssh into our host and fire up psql. I got laggy input and a bit of a startup-delay due to jumping through our bastion, but it worked well enough that I found myself staring at that no-frills psql prompt a lot.<\/p>\n\n\n\n But I kept thinking it could be better. I was right. After a few hours of ChatGPT and self-directed browsing and experimentation, I ended up assembling a surprisingly nice terminal-based workflow using:<\/p>\n\n\n\n The result feels very \u201cclose-to-hand\u201d: fast startup, pleasant interaction, integrated with the terminal environment, and has opened up a powerful cross-database workflow.<\/p>\n\n\n\n PostgreSQL already has a built-in mechanism for naming database connections: \n
pgcli<\/code><\/li>\n\n\n\npspg<\/code><\/li>\n\n\n\nduckdb<\/code><\/li>\n<\/ul>\n\n\n\nPostgreSQL Service Definitions<\/h2>\n\n\n\n
~\/.pg_service.conf<\/code><\/p>\n\n\n\n