Closed phoxelua closed 8 years ago
Hmm maybe hold back on this until we know our use cases better. It doesn't hurt to stick with pg for now?
I think we could just stick with Postgres for now, but if we want to eventually switch to Mongo, we should just do it now while we literally have nothing and are building everything from scratch.
I'm unfamiliar with the pros and cons of each, but it seems like either can work for us since we don't really even store that much data that would benefit from structure and atomic transactions (where a regular RDBMS would shine), nor do we have tons of data with high concurrency (where NoSQL might shine). I'll defer to you on this; close the issue if you don't think we should switch.
Actually yeah, we should do this sooner rather than later because SQLAlchemy doesn't work with Mongo, so we'd have to redo a lot of the models and whatever shit we set up there.
Ok I will pick this issue up next
The overall consensus is that Postgres is just superior to Mongo- in terms of memory usage, performance, and overall maturity. While our data is too small for these factors to really matter, I still would feel more at ease using something tried-and-true.
The only reason I wanted to use NoSQL was because I hate dealing with migrations....which is not particularly a good reason. Also, the fact that I suck at optimizing queries and we might kill ourselves with bad queries. But we have to learn and our models shouldn't be too complex...probably.
This will die and live on in #23
Sounds good, fuck Mongo.
Connect service to MongoDB