I’m seriously very surprised to still see so many relational databases in the top responses… guess I’m just in fantasy land. I hate writing SQL… good at it, but it’s not fun (to me).
API gateway all the things… it’s annoying at times, but I don’t want to have to manage a dbs… or write SQL.
Of course, there’s always some data store to manage, but I prefer the ones with fewer switches and easier scaling (like DynamoDB).
In the end, it’s a matter of preference. I prefer writing custom map reduce functions, you prefer SQL indexes.
Honestly, the amount of requests and data I’m handling for personal projects could be easily handled by the filesystem… so, ddb is great and I stay free tier… like always.
Use an ORM. With NoSQL on relational data, you’re guaranteed going to write a worse, relational solution than what relational databases provide. SQL is shit, but rewriting relational logic is worse.
Except when you’re not worried about scale because you’re building a small side project… I don’t want to pay for a db (or the hardware to host it) for my play projects. My in-home play server is a very old home PC that is very underpowered for today’s software.
Ah, yes… good old sqlite… I guess the main reason I go the route I go is that I like writing map/ reduce functions… much more than reorganizing tables because I structured them poorly to start.
In fact, the project I’m on right now at paid job has a lot of structure transformation… and I’m enjoying it so much that I’m not even pining for one of my million side projects.
While it’s not the most performant way to do things, I feel like data structure manipulation is one of the easiest to read ways to get from point a to point b.
I’m seriously very surprised to still see so many relational databases in the top responses… guess I’m just in fantasy land. I hate writing SQL… good at it, but it’s not fun (to me).
I hate when I have to work with non SQL based persistence.
To each their own I guess. 😇
API gateway all the things… it’s annoying at times, but I don’t want to have to manage a dbs… or write SQL.
Of course, there’s always some data store to manage, but I prefer the ones with fewer switches and easier scaling (like DynamoDB).
In the end, it’s a matter of preference. I prefer writing custom map reduce functions, you prefer SQL indexes.
Honestly, the amount of requests and data I’m handling for personal projects could be easily handled by the filesystem… so, ddb is great and I stay free tier… like always.
Use an ORM. With NoSQL on relational data, you’re guaranteed going to write a worse, relational solution than what relational databases provide. SQL is shit, but rewriting relational logic is worse.
Except when you’re not worried about scale because you’re building a small side project… I don’t want to pay for a db (or the hardware to host it) for my play projects. My in-home play server is a very old home PC that is very underpowered for today’s software.
Not sure I get what you’re saying. Sqlite exists and works on very underpowered hardware.
Ah, yes… good old sqlite… I guess the main reason I go the route I go is that I like writing map/ reduce functions… much more than reorganizing tables because I structured them poorly to start.
In fact, the project I’m on right now at paid job has a lot of structure transformation… and I’m enjoying it so much that I’m not even pining for one of my million side projects.
While it’s not the most performant way to do things, I feel like data structure manipulation is one of the easiest to read ways to get from point a to point b.