MongoDB without much trouble
Oh, how many copies are still broken - SQL or NoSQL... But still in some cases MongoDB does a much better job than MySQL or PostgreSQL. For example, the real case I was witnessing - gathering and processing hosting statistics (CPU, i/o, memory load, etc.) - MySQL did not cope with the word at all. MongoDB coped without much trouble. The database reaches the volume of 200-300 Gbytes, the data stream reaches 100 and more MB/sec. In my opinion, it is indicative.
MongoDB for you, if:
- - you don't have a clear, pre-described data structure or you assume that the data composition can then change a lot (of course, you can do it in SQL, but you have to think about other concepts, you can change it, but the question is how laborious it will be);
- - you have quite a serious amount of data planned (tens or even hundreds of GB), you can determine this even at the TOR stage;
- - you just want NoSQL, it's fashionable;
- - easy to install and try, works fine with no special settings. And if you go deeper, learn, you can set up a lot.
There are also disadvantages.
- - There are no simple transactions. At least in the classic form, as they are in MySQL/PostgreSQL. When adding a lot of data, which depend on each other, there may be certain difficulties, which will have to solve themselves at the code level. I mean, you may not encounter, of course, but...;
- - data connectivity is almost non-existent. It comes to mind at once, constantly mentioning JOIN from SQL - this is https://www.sqlsplus.com essentially not the case. Although, to be more precise, you should just think in other categories;
- - you have to rebuild your thinking to NoSQL. Otherwise, it will be difficult to maintain - that is, a good programmer database performance (or rather, a software architect) must in the future.