NoSql vs Relational DBMS. How to choose the right DBMS, so that later it would not be excruciatingly painful?

Question

SQL

As part of the project, I choose a DBMS. At the initial stage of the project, it is supposed to store these documents, which, in theory, tilts the scales in favor of NoSQL, but then a little bit of mathematics, turnover, balances, etc. will appear. In the long run, a web service that will be synchronized with the application.
What is better to choose, which then was not excruciatingly painful in finalization (development)? Relational tables or NoSql?
a) for desktop?
b) for the site?

Answer 1

Until you have a highload - choose what you know best. Believe me, it's better to work with a familiar tool than to fill bumps in a new one
. Of my preferences:
- If you are supposed to structure data and links, then SQL is without options
- If you have basic analytical queries, then SQL, but if the data becomes overdone, then look at NoSQL ColumnBased
- For cached NoSQL Key-Value data
- If you have a microservice architecture, then I would advise you to look more at NoSQL DocumentType

Answer 2

these documents

Do you understand that this is a very vague concept? The fact that in the same monge they operate with the "document" entity does not say anything for making a decision in favor of the monga :)
In the same postgres, JSONb was already brought in a long time ago and mathematics can be done there.
I studied this issue for myself and also looked towards the NoSQL database, I have data in which, in principle, there are many unknown types and their number (financial analytics), as a result, I pulled it all onto a relational model.
But in what NoSQL came in handy, it's like a cache of analytical results - so as not to load postgres with the same queries.

Answer 3

Use the SQL database with which you are familiar, i.e. you can set up the config, tweak it, measure the performance and everything will be fast. At the expense of NoSQL - in general, a dubious operating time, as for me, everything is possible in SQL, with a prepared statement, if there is a lot of time for parsing SQL (only by test results).

Answer 4

Chewed where who has the advantages:
Postgres vs Mongo / Oleg Bartunov