smith.ray•2mo agoMost Upvoted

Tried MongoDB for my first project and ended up switching back to PostgreSQL after a week

I spent 3 days fighting with nested document queries on a simple blog app before I realized the data needed to be relational anyway. PostgreSQL took me 2 hours to set up the same thing and it just works. Has anyone else jumped into a hyped tech stack only to find a basic tool was better?

3 comments

3 Comments

margaretshah2mo ago

lol oh man, your experience sounds exactly like mine from last year. I picked MongoDB for a little hobby project thinking it would be so much faster, but ended up completely lost in a sea of embedded arrays trying to join user profiles with comments. Spent a whole weekend figuring out why my aggregation pipeline was returning nothing, then realized I just needed a simple foreign key relationship. Postgres had it all working in like 20 minutes after that, so funny how the "old school" tools are often the best for normal stuff.

andrew_miller902mo ago

Yo, did you catch that talk at the last Postgres conference about how MongoDB is basically just a fancy JSON store now? I swear, everyone I know who tried that nosql hype train ended up going back to relational databases for most stuff. It's wild how people forget that joins have been a solved problem for like 40 years. Your aggregation pipeline nightmare is exactly what I hear from devs all the time - they spend days wrestling with a tool that's supposed to be "simpler" when a basic SQL query would have done the job in five minutes. Postgres just works man, no need to overthink it.

michael_bennett1121d agoMost Upvoted

3 million daily active users on MongoDB Atlas last quarter says people are finding value in it somewhere. I get that joins are elegant but sometimes you don't want your data normalized across 12 tables just to serve a single page. The document model can be faster when you actually design for it instead of trying to force SQL patterns on it. And honestly Postgres's JSON support is catching up but it still doesn't handle deep nesting as cleanly as MongoDB does out of the box. The aggregation pipeline is actually pretty powerful once you stop fighting it and learn the stages properly.