When should I use NoSQL?

Not in any scenario, NoSQL is superior to relational databases. Let's talk a little bit about when to use NoSQL to compare power:
1) database table schema changes frequently
For example, the online store, the maintenance of product properties often to increase the field, which means that the ormapping layer of code and configuration to change, if the table's data volume million, the new field will bring additional costs (rebuild index, etc.). In this scenario, NoSQL applications can greatly increase the scalability of the DB, and developers can focus more on the business layer.



2) database table fields are complex data types

For complex data types, such as SQL Sever provides extensibility support, such as fields of XML type. Many used students should know that the field, whether it is a query or change, the efficiency is very general. The main reason is that the DB layer is difficult to index XML fields efficiently, and the application layer also makes the parsing transformation from character stream to Dom. NoSQL is stored in JSON, providing support for the original ecosystem, which is much more efficient than traditional relational databases.



3) High Concurrent database requests

This kind of application is common in Web2.0 's website, many applications for data consistency requirements are very low, but the relational database transaction and large table join instead become "performance killer". In the case of high concurrency, the performance comparison between SQL and No-sql has been controversial due to differences in environment and perspective, not that in any scenario, No-sql is always faster than SQL.

4) distributed storage of massive data

Massive data storage If you use large commercial data, such as Oracle, the cost of the entire solution is very high and costs a lot of money on hardware and software. NoSQL distributed storage, which can be deployed on inexpensive hardware, is a highly cost-effective solution. MONGO's auto-sharding has been applied to the production environment.


It's not that NoSQL solves everything, like ERP systems, BI systems, and, in most cases, traditional relational databases. The main reason is that the business model of such a system is complex, and using nosql will result in increased maintenance costs for the system.

When should I use NoSQL?