At the beginning I used MySQL but very quickly I got drunk it was too strict we had to provide columns and sometimes for data i put null π values because they are completed in the future and visually I didn't like it.
Then I discovered Redis, a really powerful cache but here again it got me drunk, the key/value aspect was not enough for me anymore.
Finally I switched to MongoDB, I directly hung on, the data had no constraint we can have a document that evolves but it got me drunk too quickly it wasn't fast enough for my uses, the data duplication aspect I don't like at all and I miss the relationship a lot.
So for Rlink I selected what I like in MySQL, MongoDB & Redis π.
SGBD | SELECTED |
---|---|
Redis | TTL β± |
MongoDB | No restriction on data |
MySQL | A data structure |
π’ Another point that was important on RLink, having a query π€ speed that meets my expectations now I make 1 million in 1 second (with the answer π₯).
Then I try to meet my needs as much as possible π, such as being able to subscribe to a data and receive a notification when it changes.π¨
Now you have to ask yourself "how that a mixture of MySQL & MongoDB ?" π€, in fact it is very simple in your collection you will find you are given but in columns as in MySQL except that there the columns are specific to the data and not to the collection.
ID | NAME | USERNAME | COUNTRY | |
---|---|---|---|---|
0 | Elon | Elionne | mail.fr | FRENCH |
ID | NAME | USERNAME |
---|---|---|
1 | Stephen | Steph |
We can see that we have two data which are in the same collection but which have not the same columns π and that does not shock, the columns are not fixed at any time we can add/delete some.